Project case study
ML - Systems · 2025
Project Performance Insights (PPI)
A Streamlit-based, LLM-powered dashboard for extracting insights, detecting risks, and tracking KPIs across complex project lifecycles using document ingestion and snapshot comparison.
Highlights
- Built an end-to-end project intelligence system that ingests documents, tracks snapshots, and surfaces executive-level insights.
- Used Azure OpenAI to generate summaries, detect emerging risks, and analyze changes across project updates.
- Designed a modular ingestion and comparison pipeline supporting structured KPIs and unstructured text.
Gallery
What
Built a project intelligence dashboard to help teams extract insights, track KPIs, and identify risks across long-running project lifecycles.
The system focuses on executive-level visibility by summarizing complex project updates and highlighting meaningful changes over time.
How
Developed a multi-page Streamlit application for viewing active project insights and historical KPI trends.
Implemented a modular ingestion pipeline supporting DOCX, Excel, transcripts (VTT), and early-stage parsers for emails, PDFs, and presentations.
Stored structured project snapshots in a SQLite database, enabling snapshot-to-snapshot comparison without relying on raw documents.
Used Azure OpenAI with project-specific prompts to generate executive summaries, detect emerging risks, and analyze cross-snapshot deltas.
Built comparison and risk-detection utilities to surface changes in scope, budget, timeline, sentiment, and KPIs.
Results
Demonstrated reliable extraction of structured project snapshots and executive summaries from heterogeneous inputs.
Enabled longitudinal analysis of project performance, risks, and trends across multiple updates.
Delivered a scalable, extensible architecture that can support richer parsers, caching, and integrations with collaboration tools.