Challenge

The customer approached Quantori to enhance their bibliography search by developing a tool for article clustering. Although they had a basic prototype on Streamlit, they needed a high-performance production system, as no existing bibliographic tools provided clustering of publications. The goal was to create a complex clustering feature, optimizing data recording for graphs, and reduce the search program's time and memory usage, which initially handled only one request at a time.

Solution

Quantori Team built a high-performance bibliographic tool using the optimized datasets provided by the customer. We developed a front-end platform with enhanced UX and built the back-end infrastructure on GCP.

Outcome

The article clustering functionality helps identify the most relevant references and core articles, with metrics for ranking authors and publications within clusters. It supports exporting search results to CSV, including separate clusters, 'seed papers,' and search parameters. The system features a user-friendly interface and provides more precise bibliographic search and filtering, delivering unique data quickly.