OSMatrix - Grid based analysis and visualization of OpenStreetMap

Oliver Roick, Heidelberg University

Details about the talk:
Recently user generated content has become a new additional source for geographic information. The most prominent example for this phenomenon is OpenStreetMap, an attempt to create a free world map solely through voluntary effort. Since the data is not surveyed by trained professionals, rigorous quality control is required in order to produce a reliable dataset.

Therefore, “hidden” information which is typically not visualized in maps might indicate areas with an increased need for quality assessments. These information include user activities, such as the number of contributors or the number of contributions, the topicality of the data, the frequency of updates or the number features in a given area.

OSMatrix is an attempt to visualize such information. By aggregating numbers of several attributes into a hexagonal grid the spatial distribution of those attributes is illustrated.

The grid is generated using the Open Source library GeoTools. The investigation area is divided into hexagonal cells and the attributes are calculated for each cell. The results are tagged with a time stamp and stored in a PostGIS database for further research and visualization.

In order to visualize the data, the database is linked to a GeoServer, which provides maps of the attributes through a WMS interface. Through a OpenLayers based Client those maps can be retrieved and portrayed.

First results have shown interesting insights. For instance, in Germany and the Netherlands the number of contributions per user as well as the number of contributing users is significantly higher than in countries like Spain or Italy.
Furthermore, clusters with a notably higher number of user activity compared to their surrounding areas were identified in Poland, which might result from data imports. Here, one can assume, that there is an increased need for quality control in that adjacent areas in order to achieve a high quality data set.

With this innovative web-based approach for the visualization of spatially distributed phenomenon based on OpenStreetMap data a new tool is available, which allows for easy visual data analysis and is able to contribute to the improvement of data quality of OpenStreetMap.

Details about the speaker:
Oliver Roick is a research assistant at the Chair of GIScience at Heidelberg University. His research activities focus on volunteered geographic information from social media.

Theme: research