Finding an attractive or best-fit neighborhood for a new resident of any city is not only important from the perspective of the resident him or herself, but has larger implications for developers and city planners. The environment or mood of the right neighborhood is not simply created through traditional characteristics such as income, crime, or zoning regulations - more ephemeral traits related to user-perception also have significant weight. Using datasets and tools previously unassociated with real-estate decision-making and neighborhood planning, such as social media and machine learning, we create a non-deterministic and customized way of discovering and understanding neighborhoods.
Can we create personalized neighborhood rankings based on user perceptions?
Our project creates a customizable ranking system for the 195 neighborhoods in New York City that helps users find the one that best matches their preferences. Our team has developed a composite weighted score with urban spatial data and social media data to rank all NYC neighborhoods based on a series of questions asked to the user. The project’s contribution is to provide a scientific and calibrated understanding of the impact that socially oriented activities and preferences have towards the uses of space.