ODIN Website

1. Foreword

The goal of Open Data Watch (ODW) is to help countries build better and more resilient national statistical systems. Such systems must focus on users’ needs. Users range from policy decision makers to researchers, students, journalists, and the public at large. At the heart of user-centered and well governed statistical systems are open data practices that link data production to user service. But what is a practical path to achieve open data and how do countries monitor their progress? This is what the Open Data Inventory (ODIN) methodology, assessments, technical engagements, and website and tools provide. After five rounds of ODIN assessments, the ODIN 2020/21 assessment offers a wealth of knowledge. This report on the 2020/21 results offers a selection of findings, but there is much more to be explored from the data available. The new ODIN website provides an easier and faster way to access data. The refreshed look and feel of the Country Profiles are a good sources of information. But perhaps the most important achievement has been our extensive engagements with countries. We have learned a lot from our collaborations with countries and by participating in the UN Statistical Commission’s Open Data Working Group. All together, we feel we now have a strong community supporting open data for official statistics. Our thanks and appreciation go to all partners for working with us and to our donors for trusting in us and supporting our operations. We look forward to continuing our work together to improve open data in all countries.

We hope you find ODIN and this annual report useful. Send us any comments and feedback that you may have via info@opendatawatch.com.

Shaida Badiee

Managing Director, Open Data Watch

2. Introducing the Open Data Inventory

The 2020/21 Open Data Inventory (ODIN) is the fifth edition compiled by Open Data Watch (ODW). ODIN 2020/21 provides an assessment of the coverage and openness of official statistics in 187 countries, a gain of 12 countries since the last assessment. ODIN monitors the progress of open data that are relevant to the economic, social, and environmental development of a country. Because these data are a public good, they should be made easily accessible to all. Without them, decision makers cannot make informed decisions and people cannot hold their governments to account.

The year 2020 was a challenging year for the world as countries grappled with the COVID-19 pandemic. Nonetheless, and despite the pandemic’s negative impact on the capacity of statistics producers, 2020 saw great progress in open data. In this report we feature the stories of eight countries that have made substantial improvements in the coverage and openness of their statistical systems.

Progress, however, was not uniformly distributed. Countries in every region still struggle to publish gender data and many of the same countries are unable to provide sex-disaggregated data on the COVID-19 pandemic. And low-income countries continue to need support for capacity building and additional financial resources to overcome barriers to publishing open data.

Making data open — as measured by ODIN — is a critical step to make data accessible and useful for decision making. However, there are additional things to be done to facilitate data use and increase impact (see the data value chain), such as creating user-feedback loops, implementing programs to improve data literacy among stakeholders, devising data communication strategies that target groups of users, and building accessible, well-functioning data dissemination platforms.

Figure 1 shows each country in the ODIN 2020/21 assessment, color-coded by their overall ODIN score. Dark green countries score the highest and red countries score lowest.

Figure 1 ODIN overall scores, 2020

0 20 40 60 80 100
10 30 50 70 90

How to use this figure: Use the (+) and (-) buttons to zoom in and out. Hover over each country to reveal the name. Click a country to see a summary of scores. Clicking the country’s name will redirect you to the Country Profile page on the ODIN website.

ODIN is an evaluation of the coverage and openness of data provided on the websites maintained by national statistical offices (NSOs) and any official government website that is accessible from the NSO site. The overall ODIN score is an indicator of how complete and open an NSO’s data offerings are. It is comprised of a coverage and an openness subscore. Coverage scores show how complete a country’s data offerings are, and openness scores show how well the data meet standards recommended by the Open Definition and Open Data Charter. ODIN 2020/21 includes 65 representative indicators in 22 data categories, grouped under the major data categories of Social, Economic and Financial, and Environmental Statistics. ODIN scores are scaled to a range between 0 and 100, with 100 representing the best performance on open data.

ODIN 2020/21 includes 22 data categories, grouped under social, economic and financial, and environmental statistics. ODIN scores are represented on a range between 0 and 100, with 100 representing the best performance on open data.

You can read the Executive Summary or continue to view the full report below.

3. Findings

The Open Data Inventory (ODIN) 2020/21 assessed the coverage and openness of official statistical data in 187 countries. The median ODIN country score for 2020 is 48.8, meaning that fewer than half the countries satisfy more than 49 percent of the ODIN criteria for data coverage and openness across all data categories. Yet, this is an increase of 7 points compared to the last ODIN assessment in 2018. The global median scores for coverage and openness in 2020 are 48.2 and 51.8, respectively.

The following sections dive deeper into these scores:

  • Progress since 2016: This section looks at trends in ODIN between 2016 and 2020.
  • Regional focus: This section highlights scores at the regional level, top performing countries in each region, and the differences in each region’s open data priorities.
  • Country stories: This section offers stories about eight countries who made great progress in open data and how they did it.
  • Income focus: This section shows an analysis of ODIN scores by countries’ income groups and where lower income groups struggle most.
  • Sectoral focus: This section presents an analysis of scores by sector with a special emphasis on health data and COVID-19 data reporting, and economic statistics.
  • ODIN-Gender Data Index: This section shows the results of an ODIN subindex created to focus on data relevant to the well-being of women and girls.

3.1 Progress since 2016

ODIN scores have been rising (see Figure 2). Average openness scores have risen steadily since 2016. Coverage scores fell in 2017, in part because of changes in ODIN methodology, but have been rising ever since. For the 187 countries included in the 2020/21 ODIN, the five elements of openness increased by 30 percent from 2016, while the five elements of coverage increased by 18 percent. Because of a significant change in country coverage, scores from 2015 are not shown in this or subsequent charts and tables.

Figure 2 Coverage, openness, and overall average scores, 2016-2020

How to use this figure: Hover over the data points to see the data values. Click the score types in the legend to remove or add them from the graph.

ODIN scores, shown in Figure 3, have consistently been higher in wealthier countries, but lower-middle-income countries have made the most rapid progress over the last five years. Since 2016, the overall ODIN scores of lower-middle-income countries have increased by 39 percent. Average scores in high-income countries have plateaued since 2018, while the scores of middle-income countries have continued to rise strongly. Of greater concern are the 26 low-income countries whose scores have risen by only 12 percent, and the three other low-income countries that could not be included in ODIN assessments because they have no functioning websites (See the Data Accessibility section).

Figure 3 ODIN scores by income group, 2016-2020

How to use this figure: Hover over the data points to see the data values. Click the income groups in the legend to remove or add them from the graph.

Progress by region is shown in Figure 4. The Caribbean region, which had the lowest average score among all regions in 2016, has made the greatest proportional change since 2016. North America, comprised of the United States and Canada, which had the highest score in 2016, is the only region whose scores fell, dropping from 74.8 to 70.4 in 2020. Canada has oscillated around the high 70s since 2016 and the US in the low 70s. Together with Northern Europe and Australia and New Zealand, this group of high-income countries showed the least improvement since 2016, although they also have the highest average score in 2020.

Figure 4 Change in average ODIN scores by region, 2016-2020 (%)

How to use this figure: Click on the arrow next to the figure title to sort the data. Hover over the bars to see the data values.

The greatest improvement in openness has come from the adoption of additional download options, but improvements in licensing or terms of use and the provision of data in non-proprietary and machine-readable file formats have also contributed to improved openness scores. Most coverage elements have improved more slowly. The exception are data at the second administrative level: the greatest improvement in coverage has come from a sharp increase in the availability of subnational data since 2018.

Table 1 Change in element scores, 2016-2020

Coverage elements Average score 2016-2020 Change from 2016 (%)
First administrative level 20.1 4
Data available last 5 years 51.9 9
Indicator coverage and disaggregation 63.4 10
Data available last 10 years 49.9 13
Second administrative level 7.4 35
Openness elements Average score 2016-2020 Change from 2016 (%)
Metadata available 48.7 5
Nonproprietary 71.9 17
Machine readable 47.3 23
Terms of use 26.9 32
Download options 34.7 176

How to use this table: Sort the data by clicking the arrows to the right of the column title.

Some of the changes shown in Table 1 are correlated, as an increase in one element often corresponds to an increase in another. For example, an increase in the score for download options is most often obtained by publishing more data in data portals (or an online dissemination platform designed to make numerical data easy to reuse). Data portals almost always make data available in machine-readable formats, so the two elements tend to move together.

In another example of elements moving together, openness scores are linked to coverage scores because openness scores are only applied to datasets that are publicly available. (If no data are found, those indicators automatically get a score of 0 for all openness elements.) The coverage elements are also closely related to each other. As more datasets are published and scores for indicator coverage and disaggregation rise, 5- and 10-year data availability should rise as well, because more years of data are now publicly available. However, scores for data available over the last 10 years have risen faster than data available over the last five years, which may indicate that countries are publishing older datasets not previously available instead of more timely data. Timeliness, measured indirectly in ODIN, is an important determinant of the value and usefulness of data. The COVID-19 pandemic has shown how crucial timely data are. Timeliness is also important for longer-term purposes, such as meeting the Sustainable Development Goals (SDG).

Terms of use are consistently the lowest scoring openness element, but scores have been improving. Terms of use should include a license that outlines how data can be used and under what circumstances. (Terms of use is a broader term for a policy that may include additional language about conditions for the use of data.) Adopting an open license is a core component of the definition of open data, and data cannot be open unless they are licensed for reuse. The only way this score can be improved is by adopting or creating a new, open terms of use with an open data license or by modifying an existing one. This is in contrast to other coverage and openness elements that can see score improvements as a result of other elements improving, as illustrated by the example of indicator coverage and disaggregation in the paragraph above. In addition, some national statistical offices with no knowledge of open data could perform well on certain elements of ODIN, but terms of use is not one as it requires specific technical capacity to draft and implement. For more information about which countries have adopted open terms or use or data licenses, see the sections: Regional Focus and Country Stories.

3.2 Regional Focus

Figure 5 Changes in regional median ODIN scores, 2018-2020

How to use this figure: Use the (+) and (-) buttons to zoom in and out. Hover over each region to reveal the name. Click a region to see the median score change.

Countries in Africa and the Caribbean made the most significant improvements

ODIN findings at the regional level show the same trend of positive improvement seen at the global level but countries in the Caribbean and parts of Africa made the most significant overall improvements.. In every region overall median coverage and openness scores increased. Figure 5 shows which regions saw the most improvement (in dark blue) and those which saw less improvement (in dark orange). Countries in Africa and the Caribbean made the most significant improvements, but many other regions showed considerable progress as well, including the Pacific Islands, Eastern Europe, and Central Asia.

On the ODIN website, users can view regional results for 21 geographic regions, defined by UN Statistics Division (M49). Table 2 combines these regions into 11 regions for this section of this report based on income levels and common score trends. Table 2 shows the region’s median score from 2020, the median regional score changes for each region since 2018, and the most improved countries in each region.

Table 2 ODIN scores, score changes, and most improved countries in 11 regions

Region Number of
Countries
2020 Median Score Median Score
Change
Most Improved
Countries
Caribbean 9 48.3 15.2 St. Lucia (+44), St. Vincent & Grenadines/Dominican Republic/Cuba (+12)
Western & Middle Africa 22 41.3 8.9 Benin (+31), Angola (+22), Burkina Faso (+20)
Eastern & Southern Africa 21 41.0 7.5 Tanzania (+23), Zambia (+21), Madagascar (+20)
Pacific Islands 6 33.3 6.5 Marshall Islands (+12), Fiji (+10)
Eastern Europe & Central Asia 15 62.9 6.2 Uzbekistan (+44), Ukraine (+21), Kazakhstan (+9)
South & Central America 19 46.8 5.7 Suriname (+21), Ecuador (+11), Brazil (+8)
Southern & Western Europe 22 62.3 4.9 Serbia (+15), Montenegro (+13), Croatia (+12)
Western Asia & Northern Africa 21 54.9 4.7 United Arab Emirates (+24), Iraq (+23), Palestine (+16)
Southern Asia 9 43.1 3.8 Maldives (+13), Nepal (+7)
South-Eastern & Eastern Asia 17 52.9 2.8 Lao PDR (+16), Philippines (+15), Indonesia/Timor-Leste (+12)
N. America, N. Europe, Australia & New Zealand 14 72.1 0.3 Sweden (+6), Ireland/Finland (+5)

How to use this table: Sort the data by clicking the arrows to the right of the column title

At least one country in each region improved their score by five points. Regions with lower 2020 median scores made some of the greatest progress, but there are some exceptions. Eastern Europe and Central Asia, and Southern and Western Europe made more progress than regions with lower median scores. Even in the combined region of North America, Northern Europe, Australia and New Zealand, where little progress was made collectively, 4 out of 14 countries increased their score by 5 points. However, the combined region’s median score was pulled downward by 7 countries that saw small, negative score changes. All 11 regions saw at least one country decrease their overall score since 2018.

Open data priorities differ by region

Though ODIN scores rose across all regions shown in Table 2, they did so for different reasons. Figure 6 below shows the median score increase for each region by whether the region’s progress was driven more by changes in data coverage or openness (coverage dominant progress vs. openness dominant progress). Coverage dominant regions focused more on the publication of more data and openness dominant regions focused on making existing data more open. Many countries did a combination of both.

Figure 6 Regional median ODIN scores increases, 2018-2020

How to use this figure: Use the (+) and (-) buttons to zoom in and out. Hover over each region to reveal the name. Click a region to see the median score increase.

Figure 6 only shows part of the story. The sections below describe specific areas of coverage and openness on which regions focused their efforts. Even in regions where progress was dominated by improvements in data coverage, sizable improvements in openness were also seen and vice versa. All score increases and decreases described below refer to median score changes since 2018.

Published more indicators: Countries in Africa (excluding North Africa) increased their coverage score by publishing more new data than any other region between 2018-2020. This generally happens when countries release new data or reports or when countries develop new dissemination platforms and are able to convert older data, often only available in PDFs or paper files, into digital formats. Publishing new datasets has the biggest impact on overall ODIN scores, because making data publicly available is the first step to opening up datasets. The ODIN openness score only measures the openness of data that are already publicly available.

Published more years of data: Even though the Caribbean’s overall progress was driven by opening up already published data, they also saw a substantial increase in coverage scores from making more recent and historical data available for already published indicators. Western Africa, Middle Africa, and the Pacific Islands also made progress in this area.

Published more subnational data: Across all regions, the lowest scores are seen in publication of subnational data. This is where progress on data coverage tends to stagnate. ODIN measures whether data are published at the first and second administrative level. (It is important to note that some data categories exclude coverage at the first or second administrative level, and small countries are not expected to publish any data at the second administrative level.) However, a few regions defied the norm. South-Eastern and Eastern Asia increased their first administrative median score by over 12 points, which is the single most important reason their coverage scores increased. The Caribbean, Southern Europe, and Western Europe also saw sizable improvements in both first and second administrative level scores. And the Pacific Islands, although they improved their first administrative level score by nearly 8 points, surprisingly lost that amount and more from their second administrative level score.

Published data in open formats: An open format is defined in ODIN as data made available in both machine-readable and nonproprietary formats. Machine-readable formats make the data easy to process by a computer and nonproprietary formats ensure there are no legal or financial barriers to accessing the data due to the file format. Many of the regions where progress was led by changes in openness focused on opening up the formats of their data. For example, the Caribbean’s score increase is largely due to open formats. They saw a median score increase of 18 points for machine-readable data and an increase of 35 points for nonproprietary data. Similarly, at a small scale, we saw openness improvements follow the same pattern in Eastern Europe, Central Asia, South America, Central America, Western Africa, and Northern Africa. And while Eastern Africa, Southern Africa, South-Eastern Asia, and Eastern Asia made greater progress in the coverage of data, they also increased their machine readability score by 18 and 12 points as two combined regions.

Disseminated data through data portals: Although ODIN does not give extra points to countries that disseminate data through data portals, ODIN does measure this indirectly through its assessment of download options. Download options measured in ODIN include bulk downloads, user-select functionality, and Application Programming Interfaces (APIs). The availability of APIs and user-select functionality is almost exclusively seen in countries that disseminate data through data portals. In general, ODIN scores for this element do not increase unless a country develops a new data portal or decides to publish more data through an existing data portal. Southern Asia and Eastern and Southern Africa led progress in these area with a 15- and 13-point median increase, respectively. The Pacific Islands saw a median improvement of 9 points, while the Caribbean, Western and Middle Africa, and North America, Northern Europe, and Australia and New Zealand saw increases of 7 points.

Published more metadata: ODIN measures whether metadata are made available alongside datasets, not by checking for adherence to any particular standard, but by looking for three key components of reference metadata: date of upload, definition of indicator, and source agency. Countries that have all three elements generally have more robust metadata. The Caribbean and the Pacific Islands made the most progress in publishing more metadata. Both regions also saw similar score increases in download options, meaning they likely published more data through data portals. Countries who publish data through data portals also tend to have more robust metadata because these systems allow countries to publish consistent metadata for each dataset.

Adopted an open license or terms of use: Adopting an open data license or terms of use can have the biggest impact on data openness because one license can apply to many datasets. Terms of use should include a license that specifies how data can be used and under what circumstances. Open licenses allow users to share and reuse data for any legal purpose (including commercially) while only requiring users to give proper attribution. Countries can increase their licensing score by either adopting a new open license or removing barriers from an existing license. In Eastern and Southern Africa, licensing scores increased by an impressive22 points. Three regions showed no increase in their median scores for licensing, but others made noticeable improvements, including South-Eastern and Eastern Asia (7 points), South and Central America (6 points), and South and Western Europe (6 points).

3.3 Country Stories

This section highlights countries that made considerable progress in their ODIN scores since 2018. Making data open and accessible – as measured by ODIN – is one of the critical first steps towards the data being used for decision making. However, there are additional actions that facilitate data use and impact (see the data value chain), such as creating user-feedback loops, programs to improve data literacy among different stakeholders, data communication strategies that target specific groups of users, and building data dissemination platforms that are easy to use.

The following stories showcase not only progress measurable by ODIN, but also information from the countries themselves about the state of open data and their efforts to improve it.

Figure 7 ODIN scores in seven countries, 2016-2020

Read each country story by selecting from the drop down

How to use this figure: Hover over the data points to see the data values. Click the score types in the legend to remove or add them from the graph.

Palestine

St. Lucia has made the most progress of any country in the five-year history of ODIN with an increase of 44 points. St. Lucia, ranked in the bottom ten of all countries in 2018, has climbed to the 51st position out of 187 countries. In 2020, St. Lucia published 66 percent of the ODIN indicators, 16 percent more than were published in 2018 and increased its openness score by nearly 300 percent.

Since 2018, statistical authorities in St. Lucia have taken significant steps to improve both their data dissemination efforts and data management processes. Through assistance from Statistics Canada (STATCAN), the World Bank, Caribbean Community and Common Market (CARICOM), and the Organization of Eastern Caribbean States (OECS), the Central Statistics Office (CSO) of St. Lucia increased its commitment to leading the integration of statistics and strengthening the national statistical system (NSS).

A summary of some of St. Lucia’s activities:

Overall, the CSO strengthened its role in a government-wide approach to data governance to ensure the availability of data to inform government programs and services, and aimed to find the right balance between openness and transparency, while protecting privacy and confidentiality.

View more information about St. Lucia’s ODIN 2020/21 results on their Country Profile.

Uzbekistan was one of ODIN 2020/21’s most improved countries, matching St. Lucia's progress with an increase of 44 points. In 2018 Uzbekistan published only 39 percent of ODIN indicators. This increased to 73 percent in 2020. Their openness score increased significantly as well, with more data published in open formats and under an open license.

Between 2019 and 2020, Uzbekistan’s national statistical office, UZSTAT, with the support of the Organization for Security and Co-operation in Europe (OSCE) in Uzbekistan and the World Bank, spearheaded efforts to develop open data in Uzbekistan.

A summary of some of Uzbekistan’s activities:

  • A Creative Commons BY 4.0 license was adopted for the UZSTAT website.
  • Training was held for ten UZSTAT staff to build the capacity of government data officers who, in turn, trained almost 400 data officers in two years.
  • UZSTAT built partnerships with other stakeholders in the data ecosystem to increase data use, including the Technological Park of Software Products and Information Technologies (IT Park) of Uzbekistan, which resulted in the Open Data Challenge hackathon, a competition for the development of ICT solutions based on open data.
  • A presidential resolution was adopted in April 2019 that authorized UZSTAT to coordinate the activities of the country’s national statistical system in the collection and maintenance of open datasets and their dissemination.
  • The legislature adopted the Concept for the Development of the Open Data Sector in the Republic of Uzbekistan for 2021-2025 that aims to help improve open data in the country, as well as eliminate problems that hinder the transparency of government bodies.
  • UZSTAT attended a virtual open data workshop hosted by Open Data Watch (ODW) to better understand their ODIN 2018/19 assessment and learn how to implement ODW’s recommendations.
  • A new National Strategy for the Development of Statistics was developed.
  • A new website for UZSTAT was created in cooperation with the Asian Development Bank.
  • Uzbekistan joined the Open Data Charter in December 2020.

Like many countries, Uzbekistan created an open data portal years ago (2015), but its functionality was limited, it housed few datasets, and the use of data was relatively low. However, with the help of key partnerships and high-level political support, UZSTAT and other members of the national statistical system were able to make dramatic advances in open data and transparency initiatives more generally.

View more information about Uzbekistan’s ODIN 2020/21 results on their Country Profile.

Tanzania increased its score by 23 points between 2018 and 2020. Their improvements were driven by efforts from the National Bureau of Statistics (NBS) to increase the openness of data previously published. However, some new datasets were published as well, specifically those that include subnational data. Openness efforts focused on the publication of an open data license and making more data available through the Tanzania Social-Economic Database, which with its built in functionality makes it easier to download data in machine-readable formats.v

A summary of some of the NBS’s activities:

  • Attended a virtual open data workshop hosted by Open Data Watch (ODW) to better understand their ODIN 2018 assessment and learn how to implement ODW’s recommendations.
  • Advised ministries, department, and government agencies on how to publish data in machine-readable formats.
  • Uploaded new and existing datasets that were previously unpublished to their data portal, the Tanzania Social-Economic Database.
  • Added weblinks to other statistics producers in government to the NBS’ website to help users find other official sources of data.
  • Adopted an Open Data License for data on its website.

Hopefully these actions set a new precedent, in sharp contrast to the 2018 amendments added to their Statistical Law that made it a crime to publish statistics without prior approval. Since 2019, those amendments have been removed.

View more information about Tanzania’s ODIN 2020/21 results on their Country Profile.

In 2020, Palestine built upon the ten-point increase they earned in 2018 and gained 16 more points for an overall score of 72 in 2020. Such an achievement was made possible by the Palestinian Central Bureau of Statistics (PCBS) publishing more data and opening up the data already published, as improvements were seen equally in data coverage and openness.

A summary of some of Palestine’s activities:

  • PCBS adopted a terms of use for its data that is consistent with a Creative Commons BY 4.0 license.
  • Released a new dissemination policy in 2019 with the key principle that official statistics are a public commodity that all people should be able to access without barriers.
  • PCBS conducted a seminar on open data with local partners, including Birzeit University, the Arab American University, and the Ministry of Communications and Information Technology on open data.
  • Developed a National Strategy for the Development of Statistics (NSDS) 2018-2022 with a focus on data openness as an objective and priority.
  • Launched an Open Data Portal.
  • PCBS developed a Data Science Initiative with the Arab American University of Palestine to build the capacity of students to use their data.
  • PCBS participated in an open data workshop led by Open Data Watch (ODW) in late 2019 to improve their open data strategy and build the capacity of its staff.

At the open data workshop in October 2019, ODW provided PCBS a long list of recommendations that could be attainable within a year and would result in an estimated score increase of 16 points, a challenge they accepted and met.

View more information about Palestine’s ODIN 2020/21 results on their Country Profile.

The United Arab Emirates (UAE) increased its coverage and openness scores by more than 20 points, and its overall ODIN score increased from 51 to 75 in 2020 by making more data available and improving elements of data openness. Many of the initiatives that led to this progress were organized by the Federal Competitiveness and Centre (FCSC), UAE’s national statistical office.

A summary of some of UAE’s activities:

  • Conducted self-assessments of the open data ecosystem and a gap-analysis to provide a better understanding the assessment results and identify opportunities to improve.
  • FCSC hosted nationwide workshops in 2020 with the support of Open Data Watch (ODW) to educate and engage 13 government entities and around 60 experts of data management in government on topics of open data.
  • Launched hackathons focused on the value of data to improve government services, which played an important role to introduce, engage, and scale community involvement in the use of data.
  • Developed the National Statistical Strategy 2021-2025. This strategy includes such goals as "Promote innovation and maximize the benefits of developments in data science and smart statistics,” which led to several activities and project including the development of the UAE Open Data Portal.
  • FCSC sponsored a government competition called the “Open Data Race” (which aligned with the ODIN methodology) that encouraged entities to share their open data on their open data portal. The inaugural race saw 13 government entities taking part, publishing more than 277 datasets in less than 50 days.
  • Created a regional hub in collaboration with the United Nations to facilitate the use of big data and data science for official statistics and SDG indicators; the sharing of knowledge on newly developed methods, algorithms, and tools; and the provision of training in the use of Big Data and data science for the community of official statisticians in the Arab region.

Generally, Open Data Watch has found many country scores stagnate in the 50-60s because it requires a comprehensive effort by the national statistical office and multiple other stakeholders in the national statistical system to score higher. However, FCSC was able to make this leap in 2020.

View more information about the UAE’s ODIN 2020/21 results on their Country Profile.

Out of all countries in Eastern Europe, Ukraine saw the largest improvement since 2018 with an overall score increase of 21 points. Although Ukraine’s ranking jumped only a few places in the region, their global ranking went from 72nd out of 178 to 36th out of 187. Progress was driven largely by efforts to make existing data more open, but new datasets were also published by their national statistical office, the State Statistics Service of Ukraine (SSSU).

A summary of some of Ukraine’s activities:

In Ukraine, a policy framework designated the SSSU to receive the funding and support needed to promote and successfully implement an open data agenda.

View more information about Ukraine’s ODIN 2020/21 results on their Country Profile.

Between 2018 and 2020 Suriname increased its overall score from 36 to 57, a 21-point increase, after having a relatively unchanged score since its inclusion in ODINin 2016. Progress was made in both data coverage and openness, as more data were published by Suriname’s General Bureau of Statistics (GBS) and made available under a Creative Commons BY 4.0 license (in the page footer).

A summary of some of the GBS’s activities:

  • Participated in an open data workshop with Open Data Watch (ODW) and created a team led by the GBS Deputy Director of Research and Planning to implement the recommendations that resulted from the workshop.
  • Addressed data gaps by publishing data from secondary sources or by identifying data that were collected, but not yet published.
  • Started the process of converting PDF files to machine-readable files.
  • Removed the registration requirement for users on their website to decrease barriers to data access.
  • Cut production costs of producing hard copies of reports by making them freely available on the website.
  • Organized the Users-Producers Seminar on January 28, 2021 with the Commission for Statistics (GBS’s governing board) that resulted in a commitment to produce a National Strategy for the Development of Statistics (NSDS) within one year.
  • Created a new website for GBS in 2018 with funding from UNICEF.

Suriname’s progress was largely initiated by their dedicated team of staff who were able to make major improvements in a short time and will likely continue this progress as they accomplish longer term goals.

View more information about Suriname’s ODIN 2020/21 results on their Country Profile.

Between 2018 and 2020, the Philippines improved their overall score by 15 points to 73, propelling them into 18th place out of 187 countries, making them the highest ranked lower-middle-income country in ODIN 2020/21. This improvement builds upon the 12-point increase they made in ODIN 2017 before their progress stagnated in 2018. Since 2018, efforts by the Philippines Statistics Authority (PSA) to increase data openness have resumed.

Here is a summary of some of the PSA’s activities:

  • Held regular meetings between PSA’s Knowledge Management and Communications Division (KMCD) and Online-Dissemination Systems Team (POST) to review ODIN assessment results and create an implementation plan. The POST team includes members from all subject matters to ensure no data sectors are overlooked.
  • Designated POST members to regularly ensure that data are published whenever they become available.
  • Conducted annual capacity building activities within PSA to accomplish such goals as the rebasing of economic indicators and converting datasets into open formats to be disseminated through their OpenSTAT portal.
  • Adopted the Philippine Statistical Development Plan (PSDP) 2018-2023, which includes “Goal 1: Adequate, timely, reliable, and relevant statistics for evidence-based decision.” As part of this goal, the PSA is exploring the use of big data and citizen-generated data as possible sources of official statistics to fill current coverage gaps.

The Philippines’ efforts to publish more open data emphasize ongoing and regular actions to ensure their improvements are sustainable.

View more information about the Philippine’s ODIN 2020/21 results on their Country Profile.

3.4 Income Focus

When comparing openness and coverage scores across income groups, the difference between median openness scores of high- and low-income countries is nearly twice as high as the difference in median coverage scores. Even though improvements in openness scores have driven most countries' progress, opening up published data remains a problem for many low-income countries.

Table 3 Median ODIN scores by income group, 2020

Income Group Overall Median Coverage Median Openness Median
High income 63.1 55.6 68.0
Upper-middle income 52.3 50.2 53.6
Lower-middle income 43.8 46.8 42.5
Low income 38.2 38.6 35.1
Difference between high- and low-income 24.9 17.0 32.9

How to use this table: Sort the data by clicking the arrows to the right of the column title.

Lower-income countries struggle most with making data open, demonstrating the need for increased financial resources and capacity-building

The biggest difference between high- and low-income countries is their ability to make data available in machine-readable formats and to adopt open data licenses, as shown in Figure 8. Many low-income countries only publish data in PDF format, which is nonproprietary but not machine-readable. When data are made available in formats that are not machine-readable, users cannot easily access and work with the data, which restricts the scope of the data’s use. Depending on how a country manages their data, converting data from PDF files to machine-readable formats can be a time-consuming, manual task. More resources to help improve data management systems could greatly improve a country’s ability to make data available in various formats with little effort.

Figure 8 Median openness element scores by income group, 2020

Openness Elements

How to use this figure: Hover over the data points to see the data values. Click the items in the legend to remove or add variables from the graph.

Terms of use is another openness element that has a large gap between high- and low-income countries. This issue, unlike machine readability, is more related to capacity building. Through Open Data Watch's engagements with countries, it has become clear that the main reason most countries lack an open terms of use or data license is because of the lack of knowledge about open data and lack of technical and legal capacity to create the license.

Lower-income countries need help with publishing and collecting data more frequently

Although the difference between high- and low-income countries’ median coverage scores is about half the difference seen in openness scores, income still is a key factor in how much data a country makes publicly available. This gap is most notable in publication frequency.

ODIN assesses coverage against five elements, two of which concern publication frequency for the last 5 and 10 years. The median score difference between high- and low-income countries is 27 points for both elements. There is also a sizable gap in the number of indicators and disaggregations published by income level (16 points).

Figure 9 Median coverage element scores by income group, 2020

Coverage Elements

How to use this figure: Hover over the data points to see the data values. Click the items in the legend to remove or add variables from the graph.

It takes money and resources to collect data, so it is no surprise that countries with lower income would score less on coverage. However, as Figure 9 shows, data availability (“Indicator coverage and disaggregation”) is not the only area for concern, but also publication frequency. Countries with fewer financial resources for statistics generally collect their statistics through surveys, which are costly to implement, and thus they can afford to do so less often. Instead of focusing solely on collecting more data through surveys, administrative data or alternative sources of data should be utilized where possible. Countries can adopt statistics from international agencies as official data, create public-private partnerships, or anonymize and aggregate administrative data to help fill in the frequency gaps. Adopting alternative sources of data will not, in most cases, create comparable data time series when combined with national statistics, but it will give users a basis of comparison when published with robust metadata.

3.5 Sectoral Focus

ODIN assessments are conducted across 22 data categories grouped into 3 major categories: Social Statistics, Economic and Financial Statistics, and Environmental Statistics. In ODIN 2020/21, Food Security and Nutrition joined the previous 21 categories. Overall ODIN scores have increased year-to-year since 2016 and the sectoral analysis finds a similar trend. Nevertheless, examining the trends and levels of ODIN’s categorical scores can reveal sectors that have the farthest to go towards open data.

As shown in Figure 10, Economic and Financial Statistics have consistently received higher overall scores in ODIN assessments, for both coverage and openness scores, resulting in an average overall score of 61, compared to Environmental Statistics at 47 and Social Statistics at 44. All three sectors have increased since 2018 and indeed since 2017. However, Environmental Statistics have increased the most since 2017 overtaking Social Statistics in 2020. This reflects a large improvement in coverage and openness scores among Environmental Statistics’ categories.

Figure 10 Average ODIN scores by major data categories, 2016-2020

How to use this figure: Hover over the data points to see the data values. Click the score categories in the legend to remove or add them from the graph.

Individual data categories differ in coverage and openness, including over time

Figure 10 shows broad improvement among major data categories, but a closer look at the scores of individual data categories in the last two ODIN rounds reveals differences. Figure 11 compares the coverage and openness scores of each data category in ODIN 2020/21 to their corresponding score in 2018, grouped by major data category. Built Environment and Health Facilities are the only data categories that experienced a decline in coverage between ODIN 2018/19 and ODIN 2020/21 and Built Environment is the only data category to experience a decline in openness scores. As we can see from examining the coverage elements for Built Environment, fewer countries reported on this data category between 2018 and 2020 and this accounts for the decline in overall coverage scores. The decline in openness scores for Built Environment was driven by a decrease in the availability of metadata that may also be a function of how many countries reported on this data category.

Figure 11 ODIN coverage and openness scores by data category, 2018-2020

Economic and Financial
Statistics
Environmental
Statistics
Social
Statistics

* Category did not exist in 2018.

How to use this figure: Click the drop-down menu to choose whether to show coverage or openness scores. Hover over the bars to see the data values.

Because both graphs are sorted from highest to lowest using 2018 scores, we can observe how their relative rankings have changed over time and how coverage and openness scores differ. For example, coverage scores for data categories vary widely even within their major category. Population and Vital Statistics had a coverage score of 62 in ODIN 2020/21, compared to Food Security and Nutrition with a coverage score of 29, a 33-point difference. In comparison, openness scores across categories and within major categories are more similar. The discrepancy between openness subscores for Population and Vital Statistics and the lowest ranked openness score of Crime and Justice is only 21. This suggests openness scores are determined more by the platform used to distribute data than characteristics of the data.

Most data categories rank differently on coverage or openness scores, with some of the biggest differences observed in data categories within Economic and Financial Statistics. Balance of Payments and National Accounts are two data categories with nearly opposite rankings between coverage and openness. Balance of Payments has the highest coverage score among Economic and Financial Statistics and indeed among all data categories but has only a middling openness score, the lowest for Economic and Financial Statistics in 2018. National Accounts, meanwhile, is among the lowest coverage scores for Economic and Financial Statistics in both 2018 and 2020, but ranks much higher for openness, driven by greater download options and machine readability than Balance of Payments, for example.

Some data categories exhibit little difference between their coverage and openness scores, in addition to little movement in their overall rankings. Figure 12 highlights the rank of data categories due to overall ODIN scores in 2018 and 2020. Population and Vital Statistics ranked highest among Social Statistics in both 2018 and 2020 overall. This likely reflects that data on population are critical to the functioning of government and some of the most foundational datasets that a national statistical system produces. Crime and Justice, on the other hand, sits last on both coverage and openness. This reflects the lack of reporting on indicators and their disaggregations — particularly on sex, age, or other demographic information — in this category. The lack of sex-disaggregated data means that crimes directed against women are often underreported and overlooked. (See the discussion of the ODIN Gender Data Index.) Crime and Justice statistics are one of the world’s data blind spots.

Figure 12 Rankings of data categories, 2018-2020

How to use this figure: Click the drop-down menu to choose which major data category to display. Hover over the data points to see the data values. Click the category names in the legend to remove or add them from the graph.

Energy is the data category with the greatest change in absolute level and rank between 2018 and 2020. It became the highest scoring Environmental Statistics category in 2020 in terms of coverage and now sits at the median value for openness among Environmental Statistics. The change in the Energy scores is responsible in large part for the increase in the broader Environmental Statistics scores, seen in Figure 10. In ODIN 2020/21, a new indicator was added to the Energy category: energy supply. Generally, the addition of new indicators is followed by a decrease in scores, but here we see the opposite. Because countries were much more likely to publish data on energy supply than the existing indicator, energy consumption, scores increased. Energy supply data are often used to analyze the economic dimension of sustainable development, while energy consumption is more often used to analyze an environmental dimension, including tracking and influencing action on climate change.

Foundational health data systems are in place but inadequate to ensure no one is left behind

There are five data categories in ODIN (Health Outcomes, Health Facilities, Reproductive Health, Population and Vital Statistics, and Food Security and Nutrition) that include measures of individual’s health and the functioning of the health system. The COVID-19 pandemic has underscored the need for high-quality, timely, and interoperable health data. Because the ODIN 2020/21 assessment was conducted during the pandemic, from May to August 2020, these scores reflect, with a lag (Due to time lags in reporting, most of the indicators reviewed in ODIN date from 2019 or earlier), the data systems available at the time to provide the crucial data on a host of health-related sectors during the pandemic.

Consistent with the general trend in ODIN scores, the coverage and openness scores for the five health-related data categories have improved from 2018 to 2020. As shown in Figure 13, Population and Vital Statistics ranked highest on coverage and openness among health-related data categories in 2020 by a significant margin. As part of a country’s foundational data systems, information on a country’s population numbers recorded by the census, along with birth and death rates, are relatively available, but with important exceptions, as Open Data Watch’s work on CRVS systems with the Centre of Excellence (CoE) at the International Development Research Centre (IDRC) and UNFPA has highlighted.

Figure 13 Average coverage and openness scores for health-related data categories, 2020

How to use this figure: Hover over the bars to see the data values.

The newly created data category of Food Security and Nutrition has the lowest score for coverage and openness among the health-related categories, which highlights potential difficulties decision makers face when tracking critical issues of human health and welfare. The Food Security and Nutrition category uses official SDG indicators to assess the availability of data on undernourishment, moderate or severe food insecurity, obesity, stunting, and wasting. A low coverage score indicates that much of the data needed to track the effects of the COVID-19 pandemic on the number of undernourished or nutritionally challenged people in the world is not widely available, which has potentially dangerous implications for the ability of domestic and external relief programs to adequately provide resources.

Countries that provide more health data, publish more COVID-19 data

The world faces an immense challenge to adequately track the short- and long-term effects of the COVID-19 pandemic on vulnerable populations, as shown both above and in recent work by ODW and Data2X. To deal with the immediate effects of the pandemic on the world’s population, we need accurate data on infections, deaths, burdens on healthcare facilities as well as tests and vaccinations. Using the availability of data on these crucial indicators, as collected by Our World in Data as of late January 2021, we can match ODIN’s three relevant data categories (Population and Vital Statistics, Health Facilities, Health Outcomes) to investigate whether statistical systems that score high in ODIN 2020/21 also have more data available to fight the COVID-19 pandemic. As Figure 14 shows, this is broadly true across all COVID-19 indicators. Evidently, health systems that adequately report on health indicators during regular times are better able to respond during times of emergency as well. Countries that record higher scores for the data category Health Facilities are better able to report on hospital and ICU admissions. Likewise, countries that score higher on Population and Vital Statistics are more likely to report on deaths and excess deaths from COVID-19. This likely reflects the greater statistical capacity that is needed both to report on Health Facilities with all required disaggregations as well as to provide information by patient type and is indicative of more robust health information systems. Excess deaths can only be calculated if very timely and historical data on deaths are available, which is correlated with much more robust CRVS systems. No matter the data category, however, as the world sees glimmers of hope in increasing vaccine availability and thinks about how to build back better, statistical systems must be a priority to ensure that the current toll of the pandemic is accurately captured and to build resilience for future emergencies.

Figure 14 Average ODIN scores for health-related data categories by COVID-19 data availability, 2020

Health Facilities
  
Health Outcomes
  
Population & Vital
Statistics

Source: COVID-19 data from Global Health 5050, 25 January 2021; ODW calculations

How to use this figure: Hover over the bars to see the data values. Click the items in the legend to remove or add variables from the graph.

Economic statistics are among the highest scoring categories in ODIN, but gaps remain

Economic and financial statistics are the lifeblood of the administrative state. No matter their political persuasion, governments have always sought to capture information on their financial systems. It is therefore not surprising that coverage and openness scores of Economic and Financial Statistics are much higher than Social or Environmental Statistics, as shown in Figure 10 above. Figure 15 shows that, like other data categories, overall scores in Economic and Financial Statistics have increased over time. The ranking of most data categories is little changed from year to year, with the notable exception of Balance of Payments. International Trade and Labor have consistently ranked first and last, respectively. International trade data, captured here by merchandise imports and exports, are often highly standardized across countries that conduct trade analyses, which may explain at least part of their consistently high ranking. Conversely, labor data as captured by the employment rate, the employment distribution, and the unemployment rate, have poorer coverage, perhaps due to non-standard reporting but also to the lack of granular data by occupation or disaggregation by sex. As noted in the discussion on Figure 11 above, we can see that Balance of Payments actually occupies the top spot for coverage scores, but occupies the bottom spot for openness, revealing that the way countries publish these critical data does not facilitate their use and reuse.

Figure 15 Average ODIN scores by Economic & Financial data categories, 2016-2020

How to use this figure: Hover over the data points to see the data values. Click data categories in the legend to remove or add them to the graph.

The ability of a country to display its own financial situation is important for domestic transparency, but it can also help a country access vital financial resources in international capital markets. To guide countries in their efforts to present data that allows banks and other creditors to evaluate their investment in a country, the International Monetary Fund (IMF) established data dissemination standards in the late 1990s. These standards, which have since been updated, consist of the Enhanced General Data Dissemination Standard (e-GDDS), the Special Data Dissemination Standard (SDDS), and the SDDS Plus. The e-GDDS is a standard of statistical production that is supposed to guide all IMF member countries that choose to engage with the process to build out quality data systems and enhance statistical capacity. SDDS is specifically for countries that engage with international capital markets and SDDS Plus adds standards on top of SDDS for countries that play a vital role in global capital markets.

Figure 16 Average ODIN scores for economic and financial data by IMF SDDS subscriber status, 2016-2020

Overall scores
Coverage scores
Openness scores

How to use this figure: Hover over the data points to see the data values. Click subscriber statuses in the legend to remove or add them to the graph.

Figure 16 shows that the ODIN 2020/21 Economic and Financial Statistics scores map out the different standards very accurately. There is no difference in the rankings of each of the elements within coverage and openness across the three groups; as Table 1 shows at the global level, data at the first and second administrative level are least available for coverage and download options and terms of use constitute the lowest ranking elements for openness. What is of interest here is that SDDS and SDDS Plus countries perform somewhat similarly on coverage, but SDDS Plus countries perform much better on measures of openness, which suggests that, because of their crucial role in the global economic system, these countries have had to make their data more open. But open data on Economic and Financial Statistics and other sectors are important not just for major global economic players like the United States. Every country can benefit from opening their official statistics to shed light on development issues, fulfill their obligation to citizens, and increase access to financing.

3.6 ODIN Gender Data Index

Gender data are data that are disaggregated by sex or that measure conditions and events that have a bearing on the welfare of women and girls. They include data on health outcomes, educational outcomes, and participation in labor markets. These data are used to identify specific needs, formulate policies to address shortcomings, and monitor impacts on women and their families. Whether a country collects and publishes open gender data is an indication of the statistical system’s capacity and the country’s dedication to achieving its commitments to gender equity.

The ODIN Gender Data Index (OGDI) is a score based on the availability of 27 ODIN indicators in 10 statistical categories shown in Table 4. Of these 27, 20 indicators require sex-disaggregation or apply only to women. Seven indicators in the Built Environment and Poverty and Income data categories that are not sex-disaggregated but have important consequences for women are also included in the OGDI. Combining the scores on these data categories yields a measure of the coverage and openness of gender statistics in a country or region. The index is the unweighted average score of the 10 category scores.

Table 4 The ten data categories in the ODIN Gender Data Index

Categories with sex-disaggregated data Average ODIN score
Population & Vital Statistics 61.4
Labor 55.6
Education Outcomes 44.9
Reproductive Health 44.5
Gender Statistics 43.6
Health Outcomes 40.3
Food Security & Nutrition 37.6
Crime & Justice 34.8
Categories without sex-disaggregation Average ODIN score
Poverty & Income 46.9
Built Environment 42.2

How to use this table: Sort the data by clicking the arrows to the right of the column title.

Scores on the OGDI have been rising, but not as fast as those for non-gender data categories

Since 2016, the median score of the OGDI has risen by 21 percent, but scores on the remaining non-gender-related categories have risen by 40 percent. This reflects the slow improvement in the Social Statistics data categories compared to most Economic and Financial Statistics and Environmental Statistics data categories. As shown in Figure 17, the average score for non-gender data categories exceeded the OGDI by almost 9 points in 2020.

Figure 17 OGDI scores and non-gender data categories, 2016-2020

How to use this figure: Click on the arrow next to the figure title to sort the data. Hover over the bars to see the data values. Click the items in the legend to remove or add variables from the graph.

Gender data categories receive lower scores because they often lack historical and machine-readable data. In 2020, the OGDI coverage score was 30 percent lower than for non-gender data categories, but the OGDI openness scores was only 9 percent lower. The differences shown in Figure 18 are greatest for the availability of data in the most recent five- and ten-year periods. The lack of historical data makes it more difficult to measure progress or identify deviations from trends that may reveal underlying causes of success or failure of programs intended to redress gender inequities. Gender data are also less likely than non-gender data to be available in machine-readable formats. Gender data categories exceed non-gender categories in the availability of subnational data, but these are the lowest scoring elements for all data categories.

Figure 18 Average ODIN-OGDI and non-gender data scores by elements, 2020

How to use this figure: Click on the arrow next to the figure title to sort the data. Hover over the bars to see the data values. Click the items in the legend to remove or add variables from the graph.

Gender data are needed to manage the COVID-19 pandemic

The COVID-19 pandemic has affected men and women differently. In most countries men have higher rates of infection and death, but women are often more affected by the economic consequences of the pandemic. The OGDI indicators capture some of these effects. Countries with better gender data systems are better able to monitor the pandemic and plan for recovery.

There were 168 countries (out of 187) in the ODIN 2020/21 assessment with data available on COVID-19 cases and deaths in the Global Health 5050 COVID-19 Sex-Disaggregated Data Tracker as of January 2021. Of these, 90 countries have provided sex-disaggregated data on COVID-19 cases and deaths and 38 more provided data on cases only (33) or deaths only (5). As of January 2021, there were 59 ODIN countries without sex-disaggregated data, of which 19 published no data at all.

Figure 19 ODIN-OGDI scores by reporting of sex-disaggregated COVID-19 data

How to use this figure: Hover over the bars to see the data values.

ODIN assesses countries production of well-defined and widely used gender indicators that are — or should be — routinely collected over many years. Are the countries that score well on the OGDI also better at reporting fast-moving indictors of the COVID-19 pandemic? The answer is yes. As shown in Figure 19, there is a 13.5-point difference in the OGDI scores of countries that provide sex-disaggregated data on both the case and death rates from the COVID-19 pandemic and those that provide neither. Countries that report only cases or only deaths fall, on average, 4.6 points below full reporters.

Monitoring a pandemic requires well-developed civil registration and vital statistics and health information systems. Because women are a majority of care workers, information on their labor force attachment and paid and unpaid work is important for mobilizing resources to respond to the pandemic. And planning for recovery should include investments in education and training based on timely assessments of education and occupational skills. All of the systems that produce the necessary data are part of a core gender data system. See the forthcoming report by Open Data Watch and Data2X for a description of the core gender data systems and financing required to support it.

Low scores on the OGDI are not limited to poor countries

The lowest scoring countries on the OGDI include countries that are rich and poor and large and small. (See Table 5.) Haiti, Turkmenistan, Anguilla, and Eswatini are among the lowest-ranked countries in ODIN 2020/21 and have the lowest scores on the Gender Data Index. Anguilla, San Marino, and St. Kitts and Nevis are small and high income but all fall in the bottom 10 of the OGDI. China, a large and increasingly well-off upper-middle-income country, ranks 8th from the bottom and 23 places lower on the OGDI than its the overall ODIN ranking.

Table 5 Ten lowest scoring countries on ODIN Gender Data Index, 2020

Country Income Group ODIN Gender Data Index (0-100)
Haiti Low 0.0
Turkmenistan Upper-middle 0.0
Anguilla High 8.4
San Marino High 9.6
Eswatini Lower-middle 10.6
Libya Upper-middle 11.6
Gabon Upper-middle 13.6
China Upper-middle 16.2
St. Kitts and Nevis High 16.9
Mozambique Low 17.7

How to use this table: Sort the data by clicking the arrows to the right of the column title.

Crime and Justice Statistics are the least available data in the ODIN Gender Data Index

ODIN includes an indicator on violence against women closely related to one of the indicators for Sustainable Development (SDG) Goal 5 on gender equality. The SDGs specify five indicators that report on forms of violence, including two specifically concerned with violence against women. If countries reported these indicators with sex-disaggregated data, they would be included in their ODIN scores. The number of countries receiving no score on one or more of the OGDI data categories are shown in Table 6. Sixty-five countries received no score in the Crime and Justice category. Crime statistics that are assessed for sex-disaggregated data include the homicide rate, rates of other crimes, and data on the prison population. Frequent and accurate reporting of crime statistics is needed to halt the epidemic of femicide and violence against women.

Table 6 Data categories least available in ODIN Gender Data Index, 2020/21

ODIN gender categories Number of countries with no score
Population & Vital Statistics 8
Labor 9
Education Outcomes 16
Gender Statistics 21
Health Outcomes 25
Poverty & Income 30
Reproductive Health 32
Built Environment 35
Food Security & Nutrition 48
Crime & Justice 65

How to use this table: Sort the data by clicking the arrows to the right of the column title.

4. Data accessibility

Improving a country's ODIN score is a critical first step to encouraging data use, but there are other aspects of data dissemination that also affect data use. Open Data Watch (ODW) has developed the Data Site Evaluation Toolkit (DSET) to identify some of the functions that an open data site should have and to provide recommendations on how to improve data sites. The DSET evaluation reviews the technical functionality of websites that host data and includes in its evaluation some aspects of open data that are included in ODIN. The DSET has been used to evaluate the World Bank Data Bank, International Labor Organization’s ILOSTAT, and national statistical office websites to improve their usability, openness, and technical functionality. The Toolkit has also been used in an evaluation with the World Bank of 22 data portals in Nepal to analyze common issues a user might encounter when accessing, finding, and using data in Nepal. While the DSET covers a range of topics from linked data to website speed that could affect data use, one of the most important for ODIN is the measurement of the availability of uptime of a national statistical website.

ODW grapples with website downtime when assessors evaluate the openness and coverage of the national statistical offices’ (NSO) websites included in ODIN. Assessors must often return to an NSO site to find a window when the site is online to perform the assessment. For this round of ODIN evaluations, some countries were not evaluated due to a long or constant period of downtime during the assessment: Sudan, Venezuela, and Vanuatu. What is just a frustration for assessors is a more serious problem for data users who cannot access critical datasets when these sites are offline.

There is not a clear industry standard for uptime of a website. High availability – or an uptime percentage of 99 percent or more – is one measure. This may sound like a very high standard, but even a site with 99 percent uptime is down for 87 hours and 36 minutes a year. Downtime can cause losses in productivity when users cannot find their data; loss of resources as IT administrators struggle to get the system back online; and a loss of faith in the NSO as a reliable data source.

ODW tested the uptime for 192 NSO websites during the 2020/21 ODIN evaluation period in May to October 2020 to evaluate the state of uptime for ODIN countries. This is a broader range of countries than were included in ODIN 2020/21 because it includes country websites that had a low uptime, even if those countries were not included in ODIN because the assessors could not access them. The research team compared these findings with an uptime evaluation that was done in 2018-2019 for a blog post on how uptime could affect data use. The results are shown in Table 7.

Although lower-income countries improved their uptime by six percent since the last evaluation, uptime for NSO websites continues to fail to meet the 99 percent threshold for high availability. Only the high-income countries studied had an average of higher than 99 percent uptime. Lower-middle-income countries had the lowest percentage of uptime with an average of 94.4 percent uptime.

Table 7 Uptime averages by income level (%)

Income level Average uptime in 2020
(%)
Uptime change since 2018
(%)
Low income 98 +6
Lower-middle income 94.4 +0.1
Upper-middle income 96.2 -0.5
High income 99.5 -0.3

How to use this table: Sort the data by clicking the arrows to the right of the column title.

Sorting the findings by regions reveals geographic differences in uptime percentages shown in Table 8. Not surprisingly, wealthier regions tend to have higher uptimes. Central Asia and Middle Africa made large improvements from their uptime scores, while Northern Africa saw the biggest percentage point decrease in uptime of 13.2 percent. This large decrease in uptime in Northern Africa is not due to a drop in uptime of all countries in the region but a dramatic drop in uptime scores from two countries in the region since the last evaluation period in 2018/2019.

Table 8 Average uptime by income level

Sub-region Average uptime in 2020
(%)
Uptime change since 2018
(%)
Australia and New Zealand 100.0 0.0
Western Europe 100.0 0.0
North America 99.9 0.5
Northern Europe 99.8 -0.1
Eastern Asia 99.7 -0.2
Polynesia 99.6 6.3
Southern Europe 99.6 0.2
Eastern Europe 99.6 1.6
Western Africa 99.5 7.5
Central Asia 99.4 14.5
South America 99.3 -0.1
Central America 98.3 1.2
Southern Asia 98.3 0.1
Southern Africa 98.2 6.2
Eastern Africa 97.8 -0.7
Melanesia 96.1 -2.3
Micronesia 96.1 -3.9
South-Eastern Asia 94.4 -3.5
Middle Africa 94.3 15.4
Western Asia 93.2 -1.6
Caribbean 91.9 -0.5
Northern Africa 85.6 -13.2

How to use this table: Sort the data by clicking the arrows to the right of the column title.

Low rates of uptime may be due to human error, security, hardware- and software-related problems, interoperability and migration issues, natural disasters, and power outages. While countries need to evaluate their own unique uptime issues, one place to start is to investigate whether a server based in a cloud network would provide better service than one maintained in the country. Major cloud networks have a higher uptime and require less technical expertise to set up, making them an effective and less technically challenging option for website hosting.

While uptime is considered a clear prerequisite for data use (you cannot use a site if it is not online), there are other factors in the DSET evaluation, such as website speed, search engine optimization, and multilingual site capabilities that can also affect data use. As ODW becomes more focused on measuring and evaluating other factors that facilitate data use, these may be investigated and reported on in upcoming ODIN reports. In the meantime, more information about the DSET evaluations is available on the ODW website.

5. Data

To download more complete ODIN datasets, go to the data page on the ODIN website. You can also use the custom scoring tool to manipulate the dataset further by using custom weighting or eliminating data categories, coverage elements, or openness elements. You can do this from the data page or by clicking the gear icon wherever you see it on the ODIN website.

6. Methodology

6.1 How to use this report

The ODIN 2020/21 annual report gives an in-depth look at the results from the most recent assessments, as well as comparisons with previous years.

Navigate: To navigate through the report, use the menu on the top left of this page. Hover over each section title to reveal subsections underneath. Click any section or subsection title to move to the beginning of that section. To navigate to another section, choose a different title from the menu. Most figures and charts are interactive and have different functionality depending on their content. A description of how to interact with each figure or table appears below each one.

Read offline: To download a printable copy of this report as a PDF, scroll to the top of the page. Before the Summary there is a link on the right side that provides a non-interactive version of the report for offline reading and printing.

Download data: To download the data used in this report, click “Data” in the menu to be redirected to the ODIN website download page. To download data for specific tables or figures, click the “Download Data” button underneath each one.

6.2 Methodology guide

View the full ODIN 2020/21 Methodology Guide. Below is a summary of the main topics.

What is the Open Data Inventory?

The Open Data Inventory (ODIN) is an evaluation of the coverage and openness of data provided on the websites maintained by national statistical offices (NSOs) and any official government website that is accessible from the NSO site. The overall ODIN score is an indicator of how complete and open an NSO’s data offerings are. The summary scores for social, economic, and environmental statistics and summary scores for coverage and openness provide a picture of the national statistical systems’ strengths and weaknesses.

What is ODIN’s purpose?

ODIN helps identify critical gaps, promote open data policies, improve data access, and encourages dialogue between NSOs and data users. NSOs and their development partners can use ODIN as part of a strategic planning process and as a measuring rod for the development of the statistical system.

ODIN provides valuable information to data users across the government, the private sector, and the public about the availability of important statistical series. In addition to the ratings of coverage and openness in over twenty topical categories, ODIN assessments record the online location of key indicators in each category, permitting quick access to over 50 indicators.

Why assess national statistical offices?

ODIN assessments begin with the websites maintained by national statistical offices (NSO) because, in most countries, the NSO is the lead agency of the national statistical system, coordinating its work with other governmental bodies that produce official statistics. If an official national data source can be accessed from the NSO’s website, it is included in the ODIN assessment. NSOs, as producers and caretakers of official statistics, have a special obligation to maximize their public benefit.

NSOs can and should be the leading advocates for and providers of high quality, official statistics to government, the public, the private sector, and the international community.

How are open data defined?

There is general agreement on the core meaning of open data. As summarized in the Open Definition, version 2.1, “Knowledge is open if anyone is free to access, use, modify, and share it — subject, at most, to measures that preserve provenance and openness.” This definition has been operationalized in the International Open Data Charter. In practical terms, open data should be machine-readable in non-proprietary formats, accompanied by descriptive metadata and export options that allow customization and bulk download, and should be free to be used and reused for any purpose without limitations other than acknowledgement of the original source. These requirements have been incorporated in the five elements of the ODIN openness assessment.

What data categories are included?

ODIN assessments review published statistics in over twenty topical categories, grouped under social statistics, economic and financial statistics, and environmental statistics. In ODIN 2020/21, there are 22 categories. The default ODIN overall score weights the three groups equally. In each category, representative indicators were selected because they are frequently needed for public policies or private initiatives and because they provide evidence of underlying statistical processes for which statistical offices are responsible. The data categories in ODIN 2020/21 are:

Social Statistics

1. Population and Vital Statistics
2. Education Facilities
3. Education Outcomes
4. Health Facilities
5. Health Outcomes
6. Reproductive Health
7. Food Security and Nutrition
8. Gender Statistics
9. Crime and Justice Statistics
10. Poverty and Income Statistics

Economic and Financial Statistics

11. National Accounts
12. Labor Statistics
13. Price Indexes
14. Government Finance
15. Money and Banking
16. International Trade
17. Balance of Payments

Environmental Statistics

18. Agriculture and Land Use
19. Resource Use
20. Energy
21. Pollution
22. Built Environment

More information about the indicators in each data category and how data categories were scored can be found on each of the category pages.

What type of indicators are assessed in ODIN?

ODIN indicators are meant to be representative of the types of data a national statistical system produces. Most indicators do not have strict definitions and related indicators can act as substitutes. The selection of indicators has been informed by many international agencies, as well as country practices and the Sustainable Development Goals.

How many countries does ODIN cover?

ODIN 2015 assessed 125 countries; ODIN 2016 assessed 173; ODIN 2017 assessed 180 countries; ODIN 2018/19 assessed 178 countries; and ODIN 2020/21 assessed 187 countries.

When and how was ODIN 2020/21 conducted?

The ODIN 2020/21 assessments were carried out between May 1 and August 15, 2020. To conduct these assessments, Open Data Watch hired and trained a group of researchers skilled in various languages to complete the first round of assessments. Following the initial assessment, each country underwent two rounds of reviews. Data published after August 15, 2020 were not used in this assessment.

What is new in ODIN 2020/21?

ODIN 2020/21 sees the addition of a new category, Food Security and Nutrition that includes three new indicators. A new indicator is included in the category Agriculture and Land Use and two new indicators are included in the category Built Environment.

Do NSOs participate in the ODIN research process?

Open Data Watch invites all countries’ NSOs to provide feedback on the datasets recorded in ODIN. Each NSO was contacted at least 3 times between March and May 2020 by email. If NSOs agree to participate, they are provided a spreadsheet with the datasets found by the ODIN team and given a month to provide feedback on those datasets. Their feedback is reviewed and incorporated into the final assessment.

What is the ODIN- Gender Index?

ODIN assessments review 20 indicators in 8 statistical categories that require sex-disaggregated data or apply only to women. These 8 categories are included in the ODIN Gender Data Index along with two more categories whose data are not sex-disaggregated but have important consequences for women. These ten data categories are equally weighted in the ODIN Data Gender Index:

Sex-disaggregated

1. Population and Vital Statistics
2. Education Outcomes
3. Health Outcomes
4. Reproductive Health
5. Food Security and Nutrition
6. Gender Statistics
7. Crime Statistics
8. Labor Statistics

Not sex-disaggregated

9. Poverty and Income Statistics
10. Built Environment

Users can construct similar measures by downloading data from the ODIN website.

6.3 ODIN’s 5-year evolution

The Open Data Inventory, the flagship project of Open Data Watch (ODW), began its first assessment in the summer of 2015. Over time, the number of countries included in ODIN has grown, the methodology has improved, and the assessment process itself has been refined. Open data is an emerging field whose principles and best practices have evolved over the years. To better reflect this evolution and the changing landscape of countries' priorities to align data with the Sustainable Development Goals (SDGs), ODIN has evolved as well. Below is a brief summary of ODIN’s evolution.

In 2015, ODIN included 125 low- and middle-income countries and assessed open data across 20 data categories. In 2016, ODIN added high-income and Organization for Economic Co-operation and Development (OECD) countries for a total of 173 countries. Quarterly publication requirements were introduced for many economic indicators. And to improve the quality and consistency of the assessments, an internal system was created to collect, validate, and review assessments. In 2017 and 2018, ODW made minor updates to the definitions for indicator, categorical, and geographic disaggregation. (See Updates in 2018 for more information.) The most significant change in 2018 was the addition of a new category for crime and justice statistics. In 2020, ODIN made additional methodology changes, the most significant being the addition of a new category on food security and nutrition and the addition of 12 countries (See Updates in 2020).

2015

ODIN included 125 low- and middle-income countries and assessed open data across 20 data categories.

2016

ODIN added high-income and Organization for Economic Co-operation and Development (OECD) countries for a total of 173 countries to highlight that open data is an important issue for all countries. Also, quarterly publication requirements were introduced for many economic indicators. And to improve the quality and consistency of the assessments, an internal system was created to collect, validate, and review assessments.

2017

Minor updates to certain indicators and definitions were made. ODIN covered 180 countries.

2018

ODW made minor updates to the definitions for indicator, categorical, and geographic disaggregation. (See Updates in 2018 for more information.) The most significant change in 2018 was the addition of a new category for Crime and Justice statistics.

2020

ODIN made additional methodology changes, the most significant being the addition of a new category on Food Security and Nutrition and the addition of 12 countries (See Updates in 2020).

Because of these changes, data from 2015 are not included in the time-series comparisons of this report. ODIN 2015 remains a record of the state of open data in 2015, but comparisons with the following years may be misleading. Comparisons of data 2016 and forward are comparable, but some methodological changes may account for small score discrepancies.

ODIN is designed as a tool that will grow over time to respond to the open data priorities of the international community. In future years, additional data categories and new datasets are likely to be added as we work with countries to better understand the barriers to and opportunities for implementing open data.

6.4 Updates in 2020

There are a few methodological changes in the 2020/21 Open Data Inventory (ODIN). The main changes are described below. For further details, please read the ODIN 2020/21 Methodology Guide.

Country coverage: The total number of countries assessed in 2020 was 187, twelve new countries compared to 2018. Sudan, Venezuela, and Vanuatu were removed from this year’s assessment because their websites experienced significant technical issues during the assessment period that made it impossible to complete their assessments. Countries added in 2020 were: Equatorial Guinea, Brunei Darussalam, Palau, Tonga, San Marino, Antigua and Barbados, Dominica, St. Kitts and Nevis, and Greenland. Yemen, Bahrain and Chad made a comeback in 2020 after being previously removed because of website issues.

Geographic disaggregation: ODIN data coverage criteria requires geographic disaggregation of data, but some data categories are excluded. For example, international trade data is excluded from disaggregation at the first or second administrative level. This year, one additional category was added to the exclusion list, Energy, which no longer requires data be made available at the first or second administrative level.

Stricter metadata requirements: Open data is not just about making data available, but also making sure that data are accessible. Countries that publish their metadata in multiple locations on their website without providing an intuitive path for users to find it will be penalized. In previous years, they were not.

New data category and indicators: A new data category, Food Security & Nutrition, has been added and includes five new indicators: prevalence of undernourishment, moderate or severe food insecurity, obesity, stunting, and wasting. Other than the addition of these five indicators, we’ve also added indicators in a few other categories: GDP (income approach), commodity production, energy supply, number of rooms or bedrooms, access to electricity, and data on housing construction materials.

New guidelines for bulk downloads in data portals: ODIN defines bulk download differently than many others, giving credit for this feature if a timeseries of a single indicator is available to download in bulk (so long as it includes all geographic and categorical disaggregations ODIN requires). Now, countries that provide expansive statistical offerings within data portals will no longer be penalized for requiring users to register to download these expansive datasets in bulk due to bandwidth or other technical issues. However, it is still best practice to inform users why registration is required for this feature and to only require the bare minimum amount (name and email) during the registration process, as well as provide immediate access.

Stable country engagement: In 2020, Open Data Watch invited every country evaluated in ODIN to participate in the assessment process. During this process, countries’ national statistical offices were invited to suggest additional datasets for the ODIN team to consider to be counted in their assessment and to provide comments on their coverage gaps. In 2018, 99 countries agreed to participate. In 2020, 100 countries agreed to participate, despite staff capacity being constrained by COVID-19. The number of countries that provided comments stayed stable at 69 countries.

Sustainable Development Goals (SDG): More SDG indicators are now accepted in place of certain ODIN indicators if the corresponding ODIN indicator is not published. To see which SDG indicators can substitute certain ODIN indicators, few each section of the Methodology Guide specific to each data category.

More robust country profile pages: Country Profiles for the most recent year now include much more information. Users can now see a further breakdown of coverage scores on the “Coverage” tab, including how many total indicators a country did not publish, how many indicators lack sex-disaggregation. The “Openness” tab now includes direct links to website’s data license or terms of use, including why it was considered not open. Under “Indicators”, users can now view all the datasets used in ODIN in a searchable interface. The “Recommendations” tab not only summarizes areas of improvements but are downloadable with the accompany data needed to act. The “Country Context” tab includes information about the country’s statistical capacity, legal framework, international data commitments, and scores from other transparency and development indexes.

Updated custom scoring tool: When users visit the ODIN website, they will see a gear icon next to certain tables or graphics. When they click the icon, they’ll be able to change how ODIN scores are calculated by eliminating certain data categories, coverage or openness elements, or even changing how scores are weighted. This can be a useful tool for users who have an interest in only certain categories or just want to compare countries based on a single element, such as data licensing.

7. Acknowledgements

The Open Data Inventory (ODIN) is managed by Jamison Crowell who also authored this report with Eric Swanson, Director of Research, Lorenz Noe, Research Manager, and Caleb Rudow, Research and Data Analyst and with support from the ODIN research team including: Tawheeda Wahabzada, Laura Batista, David Benko, Samantha Copeland, Miles Johnson, Niklas Jutting, Chandrika Kaul, Erica Ness, Suzan Osman, Dominic Scerbo, Manikiran Soma, Mama Sow, Sam Stalls, Sarah Waggoner, Riley Zecca and inputs from ODW team including: Deirdre Appel, Shaida Badiee, Elettra Baldi, Martin Getzendanner, Reza Farivari, Amelia Pittman, and Alyson Marks (SDSN TReNDS). We are grateful for financial support from the William and Flora Hewlett Foundation.