NYC Bike Rush — Spatio-Temporal Mobility Analytics

Analyzing New York City bike-share demand patterns during rush hours using spatial databases, census boundaries, and interactive visualizations.

---

Project Overview

This project explores when and where bike demand spikes across New York City during rush hours. Using real-world bike station and trip data from September 17, 2024, I built a spatial data pipeline that maps trip patterns, links them to NYC census tracts, and surfaces actionable insights for smarter bike repositioning.

Key question answered: Which neighborhoods face the highest bike demand during morning and evening rush hours — and how can operators respond strategically?

---

Datasets Used

Dataset	Description
stations.csv	Bike station IDs with latitude and longitude
trip_data.csv	Individual trip records with start/end times, bike type, and station IDs
NYC 2020 Census Tract Boundaries	Official boundary file from the US Census Bureau containing borough names and tract IDs

Why Census Tracts? They provide standardized, consistent geographic units that make it easy to compare demand across neighborhoods — more reliable than informal neighborhood names.

---

Tools & Technologies

Tool	Role
PostgreSQL	Database for storing and querying tabular data
PostGIS	Adds spatial capabilities to PostgreSQL (geometry, distance, joins)
QGIS	Interactive map visualization and time-based animation
Python	Data processing and chart generation (Pandas, Matplotlib)

---

What I Built

Database & Spatial Setup

• Created stations and trip_data tables with geometry columns
• Enabled the PostGIS extension and reprojected all coordinates to UTM Zone 18N (EPSG:32618) for accurate distance calculations in New York

Spatial & Time Analysis

• Grouped trip start times into 30-minute intervals using DATE_TRUNC() and EXTRACT()
• Used ST_Within() to spatially join each bike station to its census tract
• Aggregated trip counts by census tract and time interval to find demand hotspots
• Identified the busiest stations and used buffer analysis to locate nearby stations that could cover overflow demand

Visualizations

• Animated choropleth map in QGIS showing how trip counts change across census tracts throughout the day
• Bar chart in Python showing trip volume per 30-minute window

---

Visualizations

Animated Trip Count Map (QGIS Time Manager)

gif.mp4

Trip Counts by Half-Hour Interval (Python)

---

Getting Started

1. Clone the Repository

git clone https://github.com/prachisarode95/Spatio-Temporal_Mobility_Analytics.git
cd Spatio-Temporal_Mobility_Analytics

2. Set Up PostgreSQL with PostGIS

CREATE DATABASE nyc_bike_trips;
\c nyc_bike_trips;
CREATE EXTENSION postgis;

3. Install Python Dependencies

pip install -r requirements.txt

4. Load Data

• Open QGIS DB Manager and load stations.csv, trip_data.csv, and nyct2020.geojson

5. Run SQL Analysis Scripts

• Execute the scripts in PostGIS_Queries/ via QGIS DB Manager

6. Generate Visualizations

python scripts/visualization.py

Or open the ready-to-use QGIS project: QGIS_Project/spatio_temporal_analysis.qgs

---

Project Structure

├── Data/
│   ├── Raw/                  # Original input files
│   └── Processed/            # CSV outputs from SQL queries
├── PostGIS_Queries/          # SQL scripts for each analysis step
├── QGIS_Project/             # Ready-to-use QGIS project file
├── Data_Visualization/       # Python-generated plots
├── Docs/                     # Project methodology and notes
└── requirements.txt          # Python package list

---

Results & Takeaways

• Mapped NYC rush hour bike demand down to the census tract level
• Identified peak demand windows and the stations that drive them
• Proposed a buffer-based repositioning strategy to move bikes proactively before demand spikes
• Produced map animations and charts ready for use in planning dashboards or presentations

---

Skills Demonstrated

PostgreSQL · PostGIS · Spatial SQL · QGIS · Python · Pandas · Matplotlib · Geospatial Data Modeling · Census Data Integration · Urban Mobility Analysis

---

Built by Prachi Sarode