Data as political: Offline to Online exclusions

Technology is neither good nor bad; nor is it neutral...technology’s interaction with the social ecology is such that technical developments frequently have environmental, social, and human consequences that go far beyond the immediate purposes of the technical devices and practices themselves. ~ Melvin Kranzberg

Following the discussions around open data, the issue of what data inclusions constitute privacy violations, especially when it comes to big data, is perhaps the biggest concern to open data advocates. To quote Kieron O’Hara, what we seek is “transparent Government and not transparent citizens”.  While the issue of data that should not be included has been getting a lot of attention, the issue of data exclusions that should be included has however been getting less attention. 

At the recently concluded Hivos’ Service Delivery Indicators Project Data for Education and Health meetings, attendees were quite vocal about the political nature of data, and warned that to treat it as a neutral thing, is to implicitly support the biases built into the data.  Where the data is being used for policy making, especially resource allocation (!), then it becomes even more important to pay attention to the motives, blind spots, capacities and indeed, values of the persons designing the data collection exercise, and the collecting and analyzing of this data. It’s not a stretch to say that marginalization and exclusions offline tend to be mirrored during the data collection process as can be seen in e.g. Kenya where data from North Eastern Kenya is often left out, with  “inaccessibility, security, expense, capacity” often given as the reasons for the non-collection of this data. As a result, where the counties wish to use data for policy, they find themselves having to use proxy data, or else carry out primary research with their limited resources to correct this gap. 

These data gaps matter.  When opening up the Kenya Open Data portal, President Mwai Kibaki said “The Government data website will be particularly useful to policy makers and business persons who require timely and accurate information in formulating policies and making business decisions.” Where there is no data, then policy is deduced from ‘experience’ and extrapolation, which does a disservice to these areas as all too often this information, is neither timely nor accurate. It’s become common, with perhaps the exception of the Central Government, to have data blank spots over some counties, especially those from North Eastern Kenya. 

In addition to data collected by state and non state actors, this also extends to crowd sourcing platforms developed to collect information from citizens via citizen reports. A cursory look at the platforms deployed for election, water, infrastructure monitoring etc. have most of the reports from citizens clustered around the major cities and towns, and as an interesting peculiarity of Kenyan data, areas that are not arid or semi-arid. It’s interesting how one can almost get a perfect match between rainfall patterns and socio-economic wellbeing in Kenya with the arid and semi arid lands tending to do more poorly than their greener counterparts. 

Other pertinent exclusions include data on Persons with Disabilities, with the attendant policy implications. Where data is not collected on say, accessible health and education infrastructure, then this could negatively impact educational and health outcomes of PWDs. To illustrate, if the data suggests that there are 20 facilities available to citizens in a certain county, if  all 20 facilities  are inaccessible to PWDs, then this number should read “0”. Where this data is not available, then the assumption is that the PWDs in the county are being served, while in truth they aren’t and this invisibility of the persons is reflected in subsequent policy actions and resources allocation. 

It is perhaps not overstating to suggest that these exclusions are seen where real world exclusions apply e.g. People living in informal urban settlements tend to be missing from the urban planning process, except as a problem that needs to be resolved. This was flagged by the mappers of Map Kibera who wanted to give visibility (existence?) to the people who live, work, worship etc. in Kibera, and was never visible in any Government maps. Kibera was a blank spot on the Nairobi map until young Kiberans created the first free and open digital map of their own community.

For policy makers, open data proponents and civil society, the implications for this are obvious. We need, when carrying out projects to examine what real world exclusions exist, map these, and see if they’re mirrored in the data that we’d like to use for policy. Only then can we say we’ve made a good faith effort in promoting ethical and equitable data use for policy.