ROUGH DRAFT authorea.com/95704
Main Data History
Export
Show Index Toggle 0 comments
  •  Quick Edit
  • Square cities: time dimension

    Problem Description

    This project is a continuation of two previous research projects:

    • Senseable Moscow “Moya Moskva” project (Kats 2012)
    • NYU CUSP Applied Data science research project (Kats 2015)

    All three projects are based on the same idea of explaining significant differences between cities stats, using foursquare venues data. This particular project, however, bounds to the the temporal dimension of data, analyzing venue creation through last 5 years for 8 cities, including New York, San Francisco, Shanghai, Mumbai, Moscow, Singapur, Kiev and Minsk.

    This research aims to explore three questions stated below:

    • Do all cities have similar “temporal behavior” on venue registration?
    • Do they perform similar behaviour in terms of vanue-category granulated timelines?

    While we are driven by scientific curiosity, This is not the research for the sake of research, as the answers to those questions, potentially, may lead us to the between understanding of the spatial-economical behaviour of cities.

    Data

    Research is based on foursquare service data on venues (locations), collected through official API with the custom data collector. Scraper was collecting all venues created at any time and still existing1 for the given location. Location at this moment is defined by the coordinate rectangular.

    Methodology

    Research will be based on different time-series analysis techniques, time-series correlation, KS2 tests and K-mean clustering.

    TimeLines Overview

    Venue creation normalised timeline

    On the general time line we can see that the general trend is similar, as most of the growth for each city happened on the range from 2010 till now. However, even without clust