Mismatching in data layers

Integrating data sets in a GIS often results in an improved understanding of the problem/phenomenon at hand. One could even say that data integration is the raison d’être of GISs; in any case, data integration certainly facilitates further analysis of the data.

In real-life projects the user often has to integrate data by:

• merging mismatched data layers
• choosing, in cases where two data sets of the same features exist, which set should be preferred (based on criteria that need to be defined);
• solving, for example, problems such as changes in administrative units (merging or splitting of areas) and matching these with data that only refer to an administrative name or code.

Moreover, there is always a need to merge non-spatial (statistical data, social behaviour data, ...) with spatial data. With volunteered geodata and with crowdsourcing (Web 2.0), data integration becomes both more tricky and also more important. In this respect, meta-data and lineage documentation are essential for proper data integration. The merging of mismatching data layers might require dealing with:

• mismatchings in area (spatial extent)
• mismatchings in level of detail (scale)
• mismatchings in projection (georeferencing)
• mismatchings in time
• mismatchings in accuracy
• mismatchings in data format/type (tabular and spatial data)
• mismatchings in purpose for which the data are being collected.