Enhancing Usability and Data Curation Through Front-End Design and Social Media Integration in the SMIDGen System

• Shukria Muhammadi, Computer Science, University of Delaware

• Matthew Mauriello, Computer Science, University of Delaware

This project enhances the usability of SMIDGen, a web-based tool for AI-assisted dataset generation, by improving its front-end design and integrating social media data. I focused on making the interface more accessible through responsive layouts, readable text, and interactive tutorials. Additionally, I incorporated data from BlueSky, a decentralized platform, using custom scripts to collect and filter relevant posts. A user study with 10–15 participants provided feedback that informed iterative design improvements. This work demonstrates how thoughtful UI design and real-world testing can improve accessibility and engagement in AI-driven research tools.