Accessibility and Smart Data: the Case Study of mPASS
Catia Prandi · 2014 · Proceedings of the 11th Web for All Conference (W4A) · doi:10.1145/2596695.2596723
Summary
This doctoral consortium paper presents mPASS (mobile Pervasive Accessibility Social Sensing), a research project from the University of Bologna that applies the concept of Smart Data to urban accessibility. The paper argues that while crowdsourcing and personal sensing generate vast amounts of data about the urban environment, this volume of information can itself become an accessibility barrier if not processed intelligently. Smart Data — information extracted from large datasets using intelligent algorithms to meet specific user needs — offers a way to make crowdsourced accessibility information genuinely useful. mPASS combines three data sources: expert assessments of urban accessibility, sensor data from mobile devices, and crowdsourced reports from users. It also integrates geo-referenced data from social platforms like Foursquare. The system provides personalised navigation services through a dual user profile: an Urban Accessibility Profile (UAProfile) that captures preferences about physical barriers and facilities, and an e-Accessibility Profile (eAProfile) that captures how the user needs the digital map interface itself to be accessible. The research sits at the intersection of three areas: accessible map rendering on mobile devices, user modelling for accessibility personalisation, and the emerging challenge of making large crowdsourced datasets accessible and usable.
Key findings
The paper outlines a research design rather than reporting final results, but presents several important contributions. The dual-profile approach — separating physical urban accessibility needs from digital interface accessibility needs — represents a thoughtful framework for personalisation that acknowledges users face barriers both in the physical environment and in the digital tools designed to help them navigate it. The map adaptation mechanism uses visual personalisation techniques to render map objects differently based on individual user profiles, improving both the accessibility and relevance of displayed information. The planned research agenda includes using machine learning to dynamically update user profiles based on observed behaviour, recognising that users' abilities and needs can change over time. The paper also identifies a critical emerging challenge: ensuring that the growing volume of crowdsourced and sensor-generated data remains accessible rather than becoming a new barrier through information overload.
Relevance
This paper highlights a challenge that has only grown more pressing since 2014: as accessibility data collection scales up through crowdsourcing and sensing, the sheer volume of data can itself create accessibility problems. The Smart Data concept — filtering and processing large datasets to deliver personally relevant, actionable information — remains central to modern accessible navigation tools and urban accessibility platforms. The dual-profile model (physical accessibility needs vs. digital interface needs) is a useful framework for any system that aims to help people with disabilities navigate physical spaces via digital tools. For practitioners building accessible mapping or navigation applications, the paper underscores the importance of not just collecting barrier data but presenting it in personalised, accessible ways that account for each user's specific disabilities and interface preferences.
Tags: accessible maps · crowdsourcing · smart data · urban accessibility · user modeling · mobile accessibility · personalization · navigation