I have a dataset of road accidents containing: location (latitude and longitude), date, time, country.

I would like to create a model to predict the existence of accidents for a given location.

My question is which attributes i can create to build this model?

  • $\begingroup$ There is a corpus of published material on this subject. Have you read any of their work, including what attributes they found informative? $\endgroup$ – EngrStudent May 9 '17 at 20:53
  • $\begingroup$ I found an example that includes accidents causes like weather and traffic but the problem that i'm using my app to collect data so there is no data for accident causes $\endgroup$ – HanenO May 9 '17 at 21:02

Querying location characteristics from the Google maps api or a package like a geopy and then adding weather characteristics would be a good place to start.

  • 1
    $\begingroup$ time of day, day of week, holiday status, vehicle type, level of traffic and driver type should also be considered. Weather is good, including lighting, wind, and precipitation might be good. $\endgroup$ – EngrStudent May 9 '17 at 20:52
  • $\begingroup$ @J_Heads Thank you for your response i will add it but my dataset will produce the predicted class (accident existence) only with value "yes" how can i add the "no" to my dataset $\endgroup$ – HanenO May 9 '17 at 20:56
  • $\begingroup$ @EngrStudent thank you for your ideas but i can't get this informations from my app , do you have idea how to get for example traffic level $\endgroup$ – HanenO May 9 '17 at 21:00
  • $\begingroup$ @honeyyy those types of questions are not really on topic here, opendata.stackexchange.com would be more appropriate. $\endgroup$ – GeoMatt22 May 9 '17 at 21:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.