I've dumped all the logfiles into a PostgreSQL database which has been spatially enabled using PostGIS. PostGIS allows for creating columns that contain geometry types specified in the Open Geospatial Consortium simple features standard which covers several useful ways to define 2D geometries. Additionally PostGIS allows one to use spatial indexes on these columns based on R-Trees (R as in rectangle) which speeds things up a lot since most of my queries often regard geometries that are within a more or less known area in relation to each other.
Since I also do analysis based on measurements in relation to road geometry I use maps from the OpenStreetMap project which describes itself as "...a free editable map of the whole world...made by people like you ". You can download the entire planet (big) or smaller regions (such as Sweden in my case) and then import them into PostGIS, great stuff.
The PostGIS developers also make a nice visualization tool for spatial data called uDig which is based on the Eclipse platform and thus fully extendable. The screenshot below shows some of our measurement points color-coded by vehicle speed and overlaid on the OpenStreetMap data.
Oh, and it's all open source...
