I've thought a fair bit about speedtrap registries, with my own history of getting speeding tickets, and I'm convinced they're useless. What's to stop the cops from adding a ton of false positives to the database?
Good point, and I thought a lot about that before even saying anything here.

First of all, false entries in the database are a problem regardless of who's putting them in there. Whether it's cops or script kiddies with nothing better to do, false data is definitely an issue.

There are some things that can help work around this...

- If the database is popular enough and has enough entries, and if the method for graphing the data is done well, the good data will overwhelm the bad data. On a properly designed map, the bad data would look like faint noise compared to the good data.

- To prevent a concerted effort to overwhelm the system with focused non-noisy bad data, authentication methods would be used that identified each user submitting data. In much the same way that other online services with lots of users try to prevent people from creating a bunch of accounts using automation. Then, each user would be limited to reporting a certain number of speed traps in a given time period, so that a single user can't fill up the system with his own reports.

Given that warning, would you drive (a) the 60mph posted speed limit, (b) the 75-80mph flow of traffic, or (c) the 90mph that you can do quite safely when it's not rush hour?
Up to you. The important thing is that that section of road would be shown as a "hot spot" on the graph. If time-of-day really is a factor, that could possibly also be indicated somehow on the graph, for instance with a different type of color.
_________________________
Tony Fabris