Using artificial intelligence to predict the risk of traffic incidents

Two months ago, Basque Government published what it decided would be known as “The First Call of the awards to foster and development artificial intelligence for 2021”.

The participant had to propose, and put into practice, ideas where Artificial Intelligence can be applied in the public administrations.   The call’s terms and conditions established that the 6 best proposals would be selected. One of them was the idea proposed by Serikat.

The contest required the 6 winners to conduct a proof of concept to demonstrate the feasibility of the submitted idea.

Our idea was to construct a three-day risk prediction system for traffic incidents in the Basque Country.

What we proposed is similar to the warning system that Euskalmet, the Basque meteorological office, uses for 3-day weather alerts, but for traffic-associated incidents this case.

And which data are we going to use to make that prediction? We are initially going to work with:

• The traffic incident log that is published on the Basque Country’s open data portal
• The road works scheduled for the coming days
• The weather data provided by Euskalmet, both the historical data and three-day forecast data
• A calendar of key dates/events for a given location, both historical data and data for the current year. We are here referring to situations such as patron saints’ festivities, fairs, concerts, holiday exodus, New Year, etc.

We have 3 months to demonstrate the feasibility and usefulness of our idea.