Most speed work asks whether drivers obey the sign. This scores 14,711 road segments in Thailand and Maharashtra against the speed a person outside the car can survive, checks the number against the law that set it, and names the office that can lawfully change it.
On the priority corridors drivers often travel below the limit, so the finding is not that people speed. It is that the limit itself was set for a road that no longer exists: a bypass posted at highway speed that informal traffic and roadside development have since grown into.
Filter the network down to this quarter's work list, open one unit as an evidence card with its map and the photograph the model read, see who can lawfully change that limit, and export a field-check list. No account, no licence.
Read itNine chapters: the sign, the three numbers one road carries, the gap at network scale, how the score is built, how it was checked, who can change it, and what could not be done.
Interactive map coloured by what the fix costs. Click a segment for its decision card.
The same instrument on the Indian network, where it returns the opposite answer.
Flagged segments rolled into fundable, named corridors. Returns are indicative, drawn from published intervention studies, not an economic evaluation of these roads.
One corridor worked end to end: what is wrong, what the photograph shows, what to do first, and who signs it.
python reproduce.py checks 44 of the 61 published figures
with no credentials and no confidential data, and names the other 17 rather than dropping them.