This project employed shared micro-mobility data from static stations to determine the land use surrounding these system stations. While land use data is readily available for cities like New York City, the primary purpose of this study is to explore the feasibility of such a model in regions with limited availability of land use data.
To what extent can citibike data be used in predicting land use characteristics? What particular aspects of micro-mobility data are most determinative in successfully predicting land use (e.g., gender of users, directions of travel, length of trip, time of day)?
We proposed a 2x2 methodology consists of two main steps: spatial conceptualization and modeling. The spatial conceptualization step involves segmenting the land into manageable units using two distinct methods: hexagon tessellation and Voronoi diagrams. In the modeling phase, we employ Geographically Weighted Regression (GWR) and k-means clustering to identify spatial patterns and relationships between land use and Citibike trip data.
