Abstract

Urban environments pose significant barriers to physical activity and public health, often due to a lack of motivation, insufficient adaptability to environmental factors, the uniformity of features in existing fitness apps, and limited opportunities for meaningful interpersonal interaction. This project proposes an AI-powered Walkability Assistant — a mobile application designed to foster healthier, more active lifestyles by addressing these challenges through personalized, adaptive, and community-oriented solutions. This project adopts two research methods—text mining and open-ended surveys—to identify latent user needs and areas for improvement in walkability applications. We conducted large-scale text mining and analyzed over two million user reviews from 26 leading walking and fitness apps using Latent Dirichlet Allocation (LDA) topic modeling. This analysis revealed seven core themes in user review: gamification and incentives, general satisfaction, device sync and technical issues, motivation and activity tracking, features and integration, fun and social play, and support and troubleshooting. In addition, the ongoing results from the open-ended survey are expected to offer deeper qualitative insights that cannot be captured through large-scale data alone. Building on these findings, the proposed prototype incorporates personalized route recommendations, real-time adaptation to environmental conditions, and gamified incentives to sustain user engagement. The Walkability Assistant aims to reduce sedentary behavior, enhance physical and mental well-being, and build inclusive urban communities through features that adapt dynamically to users and their environments. By uniting scalable AI analysis with human-centered design, this project contributes a robust, adaptive tool for promoting public health and sustainable urban mobility.

Keywords

Artificial Intelligence, Walkability, Public Health, Urban Mobility, User-Centered Design

Date of this Version

7-28-2025

Download

Share

COinS