Idea

Washington D.C. has many interesting things to be known. Visitors would like to receive advise where to go on Washington D.C., based on some preferences like coffees, museums, shopping, etc. Using the Foursquare API, a search for neighborhoods to visit has been built. This application works clustering and ranking neighborhoods based on similarity with user preferences on things to do. Moreover, for the more similar neighborhoods, the application will show recommendation of top venues to go while visiting a neighborhood.

Data sources

Neighborhood labels. A dataset prepared for the DC Office of Planning (DCOP) has been used

Foursquare API: A dataset of venues is built querying explore and venue endpoints in the Foursquare API.

Washington D.C. neighborhoods

The neighborhood dataset is composed by 8 attributes and 131 records. Only attributes X, Y, and NAME were used in the analysis.

Foursquare queries

Using a radius of 1000 and a limit of 500, venues for each neighborhood were obtained from Foursquare API.

In the next step, a cross-table was created indicating how many venues there are for each category in each neighborhood.

Search and clustering function in action