Data Study on Best Neighborhoods for Happy Hour
The skills the author demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
The purpose of this data study was born out of a personal desire to see which neighborhood had best access to happy hour. I was not happy with the option available to me to do such a search, and quite frankly I was frustrated. Given the tools I picked up in this part of the bootcamp I thought it would be a good way to flex my knowledge of webscraping and to expand on ideas I had gotten from speaking with my other fellows regarding their own ideas for projects.
The first thing that comes to mind when people are looking for places regarding food are typically Google and Yelp. Google's algorithm attempts to suggest locations and will recalculate points of interest based on location. In contrast, Yelp's design is based on listing much like the outdated Yellow Pages of old while only providing a small mini-map to view all the postings.
One method seeks to take the decision making out of a user's hands, and the other used a dated form of business listing with limited spatial information.
|CartoDB||allows developers to use data to create custom maps|
|Scrapy in Python||a special toolkit that enables the scraping of a series of web pages|
|MongoDB||allows for scalability of storing information in case there is more data to be collected for future work|
|ggmap in R||to convert addresses to geospatial data|
Data on The Map
note: I opted to use less filter features for the sake of cleanliness and instead chose only price range and neighborhood selection as clickable filters for my map (please click the map for to see and use the actual app)
Future Work: Though I initially only intended this map to assist me make decisions about which neighborhoods were close to happy hour locations, it could also be expanded to assist in rental searches and even real estate development decision making. As we delve into machine learning I wonder if, given points of interest and their geospatial location, I could make a webapp that would allow a potential renter to give a training set of desirable locations and then to give a set of potential rental houses and allow the computer to assist in the decision making process.