When the government released chunks of data in  under the Kenya Open Data platform, the objectives, amongst other things,  was to spur development of applications that would utilize these data to’ improve transparency; unlock social and economic value; and build Government 2.0 in Kenya’. The release of these data sets, as anticipated, has led to the development of applications that are powered by the data spanning various sectors of the economy, including health, public expenditure, education and infrastructural development amongst other things.

The momentum in the development of apps is however slowing down with each passing day as the data becomes more and more outdated and challenges in using the data cropping up. There may be many reason for the slowing pace, but presumably one of the reason could be that so far the data has been used in developing apps that are forward-looking (solving current problems), as opposed to backward looking (investigating past relationships). Old data is not necessarily a bad thing, in fact data can be a useful base in interrogating and refining our request for data (I suspect that this might have been one of the reason for the request dataset page on open Data portal).

I am sure there are quite a number of applications that are using the data sets from OpenData portal, however, there are considerably lesser visualization and analytics projects using this data. It is undisputed that the Kenya Open Data portal is a treasure trove of data but I believe that its full potential can only be realized when the data can be presented in formats and interesting relationship visualized, which will in turn make ease interpretation, and expand its usage.

Having worked with the open data for a while, we have noted a few things that pose considerable challenge to developers and analysts.

1. The shift to county administrative units from the provincial system has made some of the data recorded before the introduction of the county system quite hard to verify. This is especially prevalent on district/division data.
2. Data entry for some of the data was done quite arbitrarily, there are multiple entries, errors in the some of the data which needs thorough refining before use.
3. There are some datasets that have considerable portions of missing records; this is especially true for financial data on constituencies.
4. Some of the datasets, checked against other publicly available data, do not match up or tend to be in accurate  which means verification is needed for some data

I suspect that the small segment of over 300 datasets in the portal that has been used is largely location base (though there are a few applications of other types). We are however optimistic that future releases will have taken into the current inadequacies to create more reliable datasets account. There are positive signs that this will happen sooner, especially with the recent cabinet approval that Kenya joins the Open Government Partnership(OGP). The World Bank has also released tons of data since last year. Early this week the World Bank managing director Caroline Anstey also held a round table meeting with developers at iHub who use the Open Data. Chris Orwa represented Doban Africa at the meeting.

Presently, albeit the greatest challenge amongst developers is to build applications that can evolve as the data evolves.