Machine Learning Based Analysis of Activity Patterns to Assess Travel Behavior in Five Boroughs of New York City

Details

• Creators:
  Allahviranloo, Mahdieh ; Kannam, Nikhita
• Corporate Creators:
  City College of New York
• Corporate Contributors:
  United States. Department of Transportation. University Transportation Centers (UTC) Program ; United States. Department of Transportation. Office of the Assistant Secretary for Research and Technology ; National Center for Understanding Future Travel Behavior and Demand (TBD)
• Subject/TRT Terms:
  Analysis Cluster Analysis Clustering Methods Decision Trees Households Machine Learning Multiday Activity Patterns Neighborhoods Neural Networks Surveys Transportation Planning Travel Behavior Travel Behavior Analysis Urban Population

  More +
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States ; New York
• Edition:
  Final Report: 2024-2025
• Corporate Publisher:
  United States. Department of Transportation. University Transportation Centers (UTC) Program
• Abstract:
  This study examines changes in urban mobility patterns in New York City between 2019 and 2022 using machine learning techniques. Analysis of the Citywide Mobility Survey data (n=85,000) reveals a reduction from seven distinct activity clusters in 2019 to four in 2022. We implement three machine learning approaches - Decision Trees, Random Forest, and Neural Networks - to predict daily activity patterns, with accuracy rates ranging from 82% to 85%. The methodology incorporates a three-tier feature analysis framework considering individual, household, and neighborhood characteristics. Results show that average daily trips decreased from 3.8 to 3.2 per person, with varying impacts across demographic groups. Women's representation in travel-heavy clusters increased to 55.1% compared to 41.5% for men, while education level and household income significantly influenced activity patterns. Personal and household features proved more effective in predicting mobility patterns than neighborhood characteristics. The analysis demonstrates the viability of machine learning applications in transportation planning while identifying demographic factors that influence urban mobility patterns. These findings can inform transportation policy and planning decisions, particularly regarding service timing and accessibility across different population segments.
• Format:
  PDF
• Funding:
  69A3552344815 ; 69A3552348320
• Collection(s):
  University Transportation Centers
• Main Document Checksum:
  urn:sha-512:bb5e381ecd0ed5b4a8bcd50bc8c0ec31dffaa0aa0119bea84e7e5ec9936e2a2248cfdc3d525aed940bc2614041b9a0ef80cc774f8b464969902f2bf134b9d2f6
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/91963/dot_91963_DS1.pdf
• File Type:
  [PDF - 1.01 MB ]