The War of Neighborhoods : Using ML Algorithms to Predict Similar Neighborhoods in Delhi

A.1 Description & Discussion of the Background

Apart from being the capital city of India, Delhi is also one of the most populated metropolitan city in the world where 16.8 million people live. It covers an area of 1,484 square kilometers. This makes its population density reach an astonishing figure of 11 thousands people per square k.m. The fact that so many people are living in so less space makes the neighborhoods rich in amenities like movie theatres, shopping centers, schools and so on.

Our subject of this study has his house located in Dabri Village, Dwarka, he loves his locality mainly because of all the amenities and facilities such as parks, pharmacies, schools, malls, shopping centers, hospitals he gets in his neighborhood.

He receives a very good job offer from a reputed company located on the opposite side of the city i.e. Connaught place, New Delhi. However, given the far distance from his current place if he decides to take up the job offer, he must relocate to the New Delhi. He is willing to take up the job offer but wants to move to a neighborhood in New Delhi which is like his current neighborhood in Dwarka.

In this project we will help our subject find a similar neighborhood in New Delhi.

A.2 Data Description

We will be dealing with the following data in our project:-

• I found dataset containing Borough and Neighborhood of Delhi on the data repository of Election Commission of India, the dataset contained the names of Borough and Neighborhood in Delhi. Here is a link to dataset dataset .
• I used python library geopy to extract coordinates of neighborhoods and saved them into a csv file on my computer. Data required some scraping for geopy to correctly give the coordinates of boroughs of Dwarka and New Delhi, since they are our points of concern. Here is a link to dataset coordinates_dataset .
• I used Foursquare API to explore the most common places of a neighborhood in form of a json file.