/flyer_distribution_routes

Divide city map into distribution routes for flyers using unsupervised machine learning.

Primary LanguageJupyter Notebook

Flyer Distribution Routes

In order to facilitate manual flyer distribution (i.e. for marketing, political flyering etc.) within a given region and with a fixed number of people, a given region should be split into equally sized clusters where eventually each cluster can be assigned to a person. The clusters should be equal in size (i.e. equal number of adresses within a cluster) such that every person gets a similar amount of work. Moreover the clusters should roughly follow cityblocks for simple orientation and hence increased efficiency within a given cluster (i.e. the clusters should not separate cityblocks which would make it more difficult to visually comprehend the boarders of a specific region when flyering for the first time).

Final solution (based on a novel, self-developed algorithm):

alt text