WELFIT: A remote evaluation tool for identifying Web usage patterns through client-side logging

• We present a method to analyze usage patterns of people with/without disabilities.
• We present a tool that combines data mining techniques on detailed usage logs.
• The proposed tool was experimented during 15 months, counting on 180 participants.
• Stream of events of people with/without disabilities differ in size and composition.
• A thorough literature review and taxonomy on evaluation tools are also presented.

Although websites evaluation tools use different data sources (e.g., Web pages, server logs, and mouse tracks), few of them support remote evaluation using detailed observational data. Without considering data that represent the user’s real interaction with the interface, usability problems and/or accessibility barriers may remain unknown. This work contributes to the field by providing a tool to identify usage patterns based on client-side event logs and by presenting event stream composition characteristics. The work results from a long-term project and the tool is now available to the community. The system records usage data during real use, identifies usage patterns, and indicates potential user interface design problems. The proposed tool was experimented, counting on 180 participants, during a 15 month period collecting data from website usage. Results obtained are promising regarding the identification of usage patterns and the characterization of event streams based on the types of events that compose them.