by Viviana Zalles
During my time as an intern at Open Data Watch, I reviewed the availability and openness of data in poor, developing countries. I wanted to get a general idea of what the data environment looks like in the developing world. I asked myself: “If I were a citizen of country X and I wanted to find data about my country, could I do so and how difficult would it be?” A plausible starting point would be the country’s official statistical website. I began my search in Africa.
To evaluate the openness of websites, I looked at the breadth (temporal and topical) of the data, their availability for download in a machine-readable format, and the lack of legal restrictions for their use. These are common elements of most definitions of “open data,” (see Eric Swanson’s “How Open are Official Statistics” on the Open Data Watch website) and they are a good stepping stones for assessing the openness of statistical websites.
I found that the range of openness of the websites I reviewed was very large. And while most countries have a long way to go in making their statistical data available, there are some examples of websites that are on the right track.
One that stood out is published by Benin’s National Institute of Statistics and Economic Analysis (http://www.insae-bj.org/). They have economic, demographic, and social data available for download in Excel format as well as PDF. The economic section covers the longest time period, providing annual data from 1991 through 2013. More recent data are available for the consumer price index, but only in PDF format. Metadata are also available in PDF format. A significant omission was the lack of licensing terms.
When it comes to breadth of coverage and timeliness, Kenya’s National Bureau of Statistics (KNBS) stands out. Not only is it one of the most frequently updated sites (http://www.knbs.or.ke), it also has a very user-friendly interface. Although this might seem like a superficial aspect of a statistical website, it is not negligible: the goal of open data is to make data freely available to everyone and anyone. Having a website that is easy to navigate is an essential part of this.
Still, there are problems with the site: the data are only downloadable in PDF format, which is not machine-readable, and they are protected by copyright. The terms of use explicitly state that the website’s content can only be used “for your own personal, non-commercial use” and that you cannot “copy or post the content on any network computer or broadcast the content in the media.” This led me to think that Kenya’s statistical agency has not embraced openness. Yet on their homepage there is a link to a “Kenya Open Data” portal (https://opendata.go.ke/), which makes freely available government data, including the 2009 census, national and regional expenditures, and a wide array of health-related statistics, all downloadable in a variety of machine-readable formats. The terms of use for this website permit unrestricted use or reuse of data, but they also include a disclaimer: “Once the data have been downloaded from the agency’s site, the government cannot vouch for their quality and timeliness. Furthermore, the Kenyan Government cannot vouch for any analyses conducted with data retrieved from www.opendata.go.ke.” What is most interesting is that Kenya’ National Bureau of Statistics is one of the partners that contributes to this portal. I can’t help but wonder why the same open data criteria used for the portal are not used for the statistical agency’s website.
Another country that is admirable from a data perspective is Burkina Faso. Burkina Faso has two websites, which together provide access to the country’s statistical data: the National Institute of Statistics and Demography (http://www.insd.bf/n/) and the National Statistics Council (http://www.cns.bf/). The first publishes economic, socio-demographic, and poverty related statistics at the national and regional levels. It is fairly up to date compared with others of its kind in the region (the latest numbers correspond to July 2013) and it is easy to navigate, but it does not allow users to download the data and the licensing terms are not explicitly stated. The second website stood out to me because it is something that I have not encountered elsewhere: it comprises a list of each sector of the government and lets you see the statistical publications released by each. This is particularly useful because often the data gathered and published by other ministries or departments are not available 0n the national statistical agency’s website. Instead of having to browse through each department’s individual website, every publication is accessible through this centralized website, which clearly benefits the users. It is a valuable way of making data accessible to the public.
These are good examples, but most of the official statistical websites I explored had one or more of the following issues: they were very out of date or did not have downloadable data or the data available for download were only in PDF format. Furthermore, most sites don’t publish their terms of use, and if they do, the terms are restrictive and often state that the data are copyrighted. While many countries have embraced the rhetoric of open data, implementation lags. What is needed to make further progress? Financial support? Maybe by showing that a robust, open data environment can have tangible economic and social benefits, more countries would be willing to invest in their statistical websites. Mutual assistance? Statistical offices can benefit from the experience of open data programs in other countries. Expert advice? International agencies and NGOs may be able to help countries overcome technical barriers. And public demand? If citizens, businesses, and public institutions tell their governments they need open access to data, statistical offices will respond.