Why OKF Global Open Data Index 2015 is a Failure

The Open Knowledge’s Global Open Data Index 2015 (GODI) edition has just been released; it covers 122 countries, which is 25 new ones from 2014. Compared to the last edition, this year the GODI is marked by the addition of four new indicators: Procurement Tenders, Water Quality, Land Ownership, and Weather Forecast.

Top 10 ODI 2015 :

The GODI found that four countries have an impressive performance: Taiwan, Colombia, Uruguay, and the Netherlands; the only country that was able to keep its position is the United States. However, the GODI found an impressive decrease in the score of France, which lost seven positions; and Norway, which lost three positions. Finally, the five other countries: UK, Denmark, Finland, and Australia fell only by one position.

Chart 1 : Yearly performance of top 10 countries in ODI 2015

Why the ODI 2015 is a Failure

For this year’s edition, four new indicators have been added. However, only 9 percent (156 open datasets from 1586) of datasets worldwide were found to be open, which is too low! This figure is a decrease when compared to 12 percent for the last edition. Thus, according to the GODI 2015, efforts have been done during 2015 that lead to a negative progress in Open Data, which is not a good approach for the GODI in my opinion.

Some research issues in the GODI are also the cause of this “impressive performance”. For example, Procurement Tenders are not available for bulk download in Taiwan, while the GODI mentions them as available. Bulk download feature seems to be unavailable in many others datasets, but it is marked as green for Taiwan everywhere. Reviewing more countries in details will certainly lead to other similar discrepancies, which proves that there are issues in the research process. A review process could help reducing these problems.

Arab countries performance in GODI 2015 :

Besides being interested in opening data in the world, I have worked specifically on open data in Arab countries, so I have lots of expertise here. However, I will not go through all countries, but I will just pick some random countries instead. For example, in Oman, the government budget is not openly licensed, but the GOBI marks it so (it is marked green).

Another example that is very clear to me is the Location Datasets which is all green in Morocco while data submitted in this category does not meet the minimum conditions, especially for the Borders polygon. Which mean all this section should be in red for Morocco.

Other datasets are totally missing, for example, this year Tunisia received the regional open government awards for their e-Procurement system TUNEPS, which contain everything procurement tenders. But the GODI 2015 only mentions the old website and, of course, this section is all in red.

In total, this year 14 Arab countries are featured in the GODI compared to only seven for the 2014 edition. I am very confused, and I cannot say whether the +27 progress in Oman’s score is due to a specific effort in any field related to opening up government data, compared with a massive fall for Saudi Arabia and for Tunisia (-29 and -20 respectively), which is very strange.

Conclusion

In my humble opinion, the GODI 2015 totally missed the point, and the result is very confusing and far from being accurate. For researchers and professionals interested in Open Data, you can still use the GODI 2015, but with caution. The only correct statement that I found is that “This year’s Index showed impressive gains from non-OECD countries”. Otherwise, the entire Index needs to be reviewed.

You can check the #GODI15 here : http://index.okfn.org/

Hatem Ben Yacoub, Independent Open Data Expert.