Open aceriverson opened 3 years ago
I like the idea of using the strava api segment explorer to search a box ([highest x, highest y] - [lowest x, lowest y]), convert the <= 10 segment titles returned to a string, plus a string of all segment titles run during the activity. Then score each location by how many times each word is in the earlier crafted segment titles string, and average by # of words in location title. If there is a tie, the closest location should win, and if there is 0 hits, the most reffed location should win.
Instead of prioritizing references of POIs, find best match based on segment titles. This will select more accurate locations and prioritize where the user ran the most. After the algorithm is in place, we can increase the search radius to 1 kilometer again.