You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Adding hotspots per types. Starting by looking at the past year of data (to avoid seasonality) and finding locations where an average of at least 1 request was made per month. Using DBSCAN to create clusters. Then finds the centroid for the points in each cluster. Tweaked the epsilon parameter to around 30m which seems to make good clusters. Want to combine points at neighboring addresses without combining those across streets. Will need to show this to users.
The text was updated successfully, but these errors were encountered:
This is pretty cool. We would need to figure out where to use this. Maybe on the map itself? Or we could also use it on a dashboard. The behavior most likely would be to pick a type and show the hotspots for it.
Description
Adding hotspots per types. Starting by looking at the past year of data (to avoid seasonality) and finding locations where an average of at least 1 request was made per month. Using DBSCAN to create clusters. Then finds the centroid for the points in each cluster. Tweaked the epsilon parameter to around 30m which seems to make good clusters. Want to combine points at neighboring addresses without combining those across streets. Will need to show this to users.
The text was updated successfully, but these errors were encountered: