Full-stack, Data Science & AI

BiblioDash

SaaS bibliometric analytics platform for universities, with interactive dashboards, institutional OKRs, and an AI assistant via MCP.

Description

BiblioDash centralizes an institution's scientific output (OpenAlex data) into a dashboard with KPIs, trends, thematic-area analysis, author/journal rankings, international collaboration, and editorial quality (quartiles, SNIP, CiteScore). It auto-generates research OKRs per area, exports PDF/Excel/Markdown reports, and offers an AI assistant via MCP on the Enterprise plan. Next.js/React frontend with NextAuth and TanStack Query; FastAPI (Python) backend with Celery/Redis for async processing, Postgres for the relational model, and Cloudflare R2 for the Parquet datasets.

Challenge

Turn large volumes of per-institution bibliographic data into actionable indicators for vice-rectors and accreditation processes, without blocking the user experience while it's being processed.

Solution

Decoupled architecture: heavy ingestion and computation run as async jobs (Celery/Redis) separate from the frontend, Postgres for relational data and R2 for the per-institution publication Parquet files; tiered subscription plans (Free/Institutional/Enterprise) gate access to institutions, exports, and the AI assistant.

Results

Live product with a tiered subscription model, dashboards with a dozen-plus visualizations, auto-generated OKRs per research area, and an AI assistant (MCP) on the Enterprise plan.