Friends of Snakes Society (FOS), based in Hyderabad of Telangana, presented a weather-based machine learning model at the First World Congress on Snakes held in Kandy, Sri Lanka, from October 1 to 4. The model helps estimate the likelihood of snake encounter during a given hour, which could help people in taking additional precautions during periods of higher encounter activity, while allowing rescue teams to prepare for increased demand, as per a press note.
Researchers developed the model using 35,234 rescue dispatches recorded in Telangana in 2024 and 2025. The records were matched with hourly weather conditions and information about sunlight, then used machine learning to identify patterns associated with snake encounters.
Avinash Visvanathan, general secretary of the society, said that they retrieved 31 weather variables and solar geometry on the days snakes were rescued by FoSS volunteers in the last two years. Details on the location and time were excluded so that the machine model can be trained exclusively on weather related model.
Apart from organisations engaged in rescuing snakes, the model can be used by government health departments to be better prepared to attend to possible increase in snake bite cases by stocking up anti-venom at healthcare centres where snake sightings are predicted to increase, said Mr. Visvanathan.
The Hyderabad-based snake rescue organisation currently puts out hourly forecasts for 72 locations across Hyderabad and surrounding districts of Telangana. The forecast for the next day are updated four times a day.
