Towards a national spatial data infrastructure (NSDI): inventory and evaluation of existing geospatial datasets and systems in Kenya

Abstract

In Kenya there are presently about 50 institutions with GIS installations and it has generally been noted that there are duplications of effort in the production of geospatial data. Again, co-operation among these institutions is minimal. In order to realize an effective, infrastructure-enabled data sharing arrangement, NSDI in Kenya, an inventory of the available geospatial datasets and systems is needed. From such an inventory, it is possible to determine the capabilities and needs of individual institutions and also to identify specific areas for increased co-operation. A study was therefore undertaken to make such an inventory, determine parameters to use to evaluate datasets for suitability and systems for compatibility and to make the evaluations. Parameters used to evaluate datasets for suitability and systems for compatibility together with the appropriate evaluation criteria were determined. The dataset parameters included: availability, accessibility, form, completeness, format, co-ordinate system, source scale, and positional accuracy among others, while system parameters included: the operating system, GIS software, data interchange formats, database architecture, Database Management System (DBMS), type of network traffic, data transfer rates, search/retrieve protocol and registration system for the SDIservers. In general, the evaluation of systems was not as elaborate as for datasets. Insufficiency of data collected during the survey limited a complete evaluation of datasets on the basis of coverage, currency, positional accuracy, price and logical consistency, while for systems the limitation was in the computer memory and speed, the database architecture, type of network traffic and rate of data transfer. Based on the evaluation of datasets, it is concluded that Kenya is about halfway ready for NSDI implementation. This is because about half of the datasets were determined as suitable. The study recommends a number of issues among them: Institutions that do not have a catalogue ()f their datasets to begin documenting their datasets; particular scales (resolutions) be agreed upon for the development of national GIS framework data; datasets not based on the Arc datum 1960 be transformed to this datum; individual institutions to prepare specifications that will guide them in the procurement of interoperable system components; and finally, for a more complete picture, another study on the other parameters be undertaken.