# Get new features from map using pickup/destination longitude and latitude to improve arrival time expectation

I am learning a model to predict the estimated time of delivery for orders. Specifically, from the instance of driver pickup to the point of arrival at its final destination.

I attempted this with some features such as distance, day_of_week, month, time, temperature, etc.

I would like to improve my model by getting new features by using existing features from maps (open_street_map, ArcGIS, or uber_movement_data)

What I would like to get from map data are two things:

1. Speed data: historical average speed data over a certain time as well as official speed limits (using vehicle_type).

2. Traffic conjunction level: between these two points

P.S I do't have the route, so I would like to get the general data between two points not the speed and state of the traffic for every single road.

• Notice, grammar/formatting corrections. Please keep this in mind for your next post. – Ethan Oct 7 '19 at 1:35