Turning 5 years of public ERC funding data (822 grants) into interactive dashboards in 15 minutes using Claude AI and Google Sheets MCP.
Data source — EC R&I Dashboard CSV export
Data pipeline — Google Sheets MCP connection
Analysis engine — Claude AI for data processing
Output layer — Chart.js HTML dashboards
Reference data — CORDIS for cross-checking PIs
Analyze 5 years of ERC grants (institutions, domains, funding) without relying on complex BI tools.
Exporting, cleaning, and visualizing the EC’s public data requires specialized skills most researchers lack.
The bottleneck wasn’t finding the data—every funded project is publicly documented. The real barrier was having the technical capacity to clean and analyze it at scale.
Unintuitive BI Tools — Applying correct filters across multiple grant types in Qlik Sense is confusing.
**Export raw CSV directly** once filters are set.
Messy raw data — Exported datasets contain inconsistent naming and nested fields.
**Pass data to Claude via MCP** for context-aware cleaning.
Slow multi-grant analysis — Different grant types require separate analysis pipelines.
Claude processed all four in a **single session**.
**Four interactive dashboards** covering STG, COG, ADG, and SyG
**Key patterns identified**: Germany/Max Planck lead STG/COG/ADG
**Time reduced from days to ~15 minutes** (export to dashboard)