Like many developing countries, access to geospatial data presents a big challenge. One of the major reasons for this is a lack of a National Spatial Data Infrastructure (NSDI) through which geospatial data collected by different stakeholders can be shared. As a consequence there is duplication of data collection efforts, usage of geospatial data with different cartographic properties for the same areas, high cost of data collection etc. Other challenges include the fact that this data is rarely updated and there is bureaucracy attached to accessing these datasets. The conglomeration of these challenges has inspired the consideration of alternative sources of geospatial data, more so freely available geospatial data. OpenStreetMap (OSM) is a geospatial dataset whose main distinguishing quality is that it is free to access and distribute. It contains information about cultural features such as roads, buildings, powerlines etc., collected through crowdsourcing. Crowdsourcing for the Uganda OSM dataset began in 2012 and to date a substantial amount of data has been accumulated, especial for the capital city – Kampala. This dataset is continuously updated through the efforts of volunteers some of whom have had insufficient training in mapping or cartography. To members of the geospatial community, this potentially presents doubts about the accuracy of the dataset and hence its usefulness. It is in this context that this paper therefore sought to assess the Kampala OSM dataset. The assessment involved comparing the OSM dataset against three known existing datasets: The Uganda National Roads Authority (UNRA) dataset, Justice Law and Order Sector (JLOS) dataset and extracted roads from a 2014 orthophoto. These were then used to assess positional accuracy, data completeness, consistency and attribute accuracy of the OSM data.

The results show that OSM data had positional accuracies of 91%, 62%, 53% when respectively compared to UNRA, JLOS and extracted road data from 2014 Kampala orthophoto. This can be explained by the fact that OSM data is mostly collected using hand held GPS which can have positional accuracies of up to 10m.  With a 99% data completeness accuracy, evidently much of Kampala has been collected which means that OSM can be reliably used for routing studies among others. There was however poor consistency and attribute accuracy of 5% and 23% respectively which could be attributed to the reference data having out of date place names. There also could have been instances where and when the wrong attribute e.g. road name was assigned to the OSM dataset. This can be remedied by ensuring proper quality control before uploading the data for sharing. From the questionnaires, it was observed that there has been growing use and application of OSM data in Geospatial community. Most awareness has come out of training workshops however, more can still be done. There is also need to expand the domain areas in which OSM data can be used, as this will go a long way in reaching out to more potential users. Ultimately as OSM takes root, potentially less resources will need to be spent on collecting this data.