How to find the nearest neighbors for latitude and longitude point on python?

Question

Input:

point = (lat, long)
places = [(lat1, long1), (lat2, long2), ..., (latN, longN)]
count = L

Output: neighbors = subset of places close to the point. (len(neighbors)=L)

Question: Can I use kd-tree for quick nearest-neighbors lookup for points with latitude and longitude? (For example, implementation in scipy)

Is it necessary to transform the geographical coordinates (latitude and longitude) of the point in the coordinates x,y?

Is it the best way to solve this?

Answer 1

scikit-learn provides a BallTree class that supports the Haversine metric. See also this SO question.