Hello, I only recently came across your fascinating project and am quite interested in contributing ideas or code to the project. I work as a GIS professional and am interested in enhancing the PwnMAP with richer details and GIS functionality. My experience is primarily with ArcGIS Pro, ArcEnterprise, and SQL Server/Postgresql, and I believe I could use these programs to create a more sophisticated geodatabase and map layout for data captured from pwnagotchis. Any georeferenced ‘pwns’ could be displayed with a fairly high degree of spatial accuracy on a map. Maps could also include additional layers that provide graphical representations of certain spatial statistics that may be of interest, such as number of devices encountered per some unit area, or areas likely dense in wifi coverage based on freely available census demographic data. There are many other freely available spatial statistics that may be of interest, and I welcome any additional input if you can think of something that could be mapped.
I also believe it may be possible to integrate demographic GIS data with your AI algorithm. For example, with gps enabled pwnagotchis, a pwnagotchi could be spatially “aware” that it was located in an area that would likely have large numbers of SSIDs broadcasting, based on it being in an area with high population density (again, from census data). This could allow its initial search parameters to be set more appropriately for a particular environment (e.g. stationary, moving; rural, urban, suburban). In theory, I could imagine this extending even to features such as highways and major roads. For example, if the device estimated that it was on a highway based on its gps location, it could set parameters appropriate for signals that are rapidly appearing/disappearing.
If you think there is any potential in these ideas, let me know and I would like to help develop them further.