Currently, the runtime efficiency is pretty bad and could most definitely be improved. The reason for this is that currently, the program goes through and scrapes each page of the results associated with the search in question and creates a list of all of the new businesses/locations that it finds through conducting this search. It then passes this entire list of (name, address) tuples into the DistanceMatrix function in order to generate the matrix through OSRM API. The problem is that within this function, we then iterate through the entire list again and geocode the addresses using geopy. There's also about 400 addresses/locations/businesses on average. Should look into either cutting down the locations to be only those that have the specific city name within their address (many of the ones returned for larger cities are located in outlying suburbs) and should look into moving the geocoding function to where the information is scraped so that all conversions take place at the same time.