Abstract

Due to the lack of a pipeline that can store and analyze a wide variety of biomedical data, our team built GILBERT. GILBERT is a web application that uses Django and stores its data in PostgreSQL. It contains four main components: a system that controls which users have access to which studies, an uploader which allows users to add studies to GILBERT's database, an interface that displays study information and metadata, and a feature that analyzes studies in the database through statistical and graphical means. Our team focused on building these components first since we believed these to be the essential elements of a data analysis pipeline. To evaluate GILBERT, twenty-seven participants responded to two surveys and attended a virtual evaluation session. In the evaluation sessions, a team member read tasks to the participant while another team member scored the participant on their behavior, their ability to complete the tasks, and understanding of the system. Overall, tasks had a high completion rate. Participants rated GILBERT as easy to use and easy to understand. Differences in ratings were insignificant between different majors, amounts of experience with large amounts of data, and experience with other data handling tools such as MATLAB, Python, and SQL. While these results show that GILBERT is an effective tool for biomedical data handling, future work can make GILBERT's uploader more flexible, allow for easier cross-study analysis, and add more options to create charts.

Download

• Complete report: pdf