Data & Methodology
How Solardash is built — the static-first model, every dataset's provenance, what's modelled, and how to refresh it.
Methodology
Static-first — no live database.
Solardash is static-first. Offline pipelines (tsx scripts) read human-maintained feeds in manual-data/, derive metrics, and write committed JSON snapshots. Pages are prerendered (force-static) and read those snapshots — there is no runtime database or client-side fetching, so every figure is reproducible and versioned in git.
Confidence legend
- HighOfficial / disclosed — MNRE, CEA, SECI, or company filings.
- MediumPublic data with Solardash aggregation / derivation.
- ModelledSolardash estimate or model (splits, forward curves, attribution).
Data provenance
Every committed snapshot — auto-generated from the live data layer.
Snapshot provenance
Section · dataset · cadence · vintage · sources
Modelled metrics
What's estimated, how, and where to refine it.
Quarterly cell production
ModelledEach producer's FY26E annual output is split across quarters with a fixed ramp [0.22, 0.24, 0.26, 0.28]; real (player, quarter) actuals override the modelled split.
Refine in manual-data/manufacturing/cell-production-quarterly-override.csv
IPP funnel splits
ModelledPer-developer operational / under-construction / pipeline and the tech mix are partly modelled from investor disclosures.
Refine in manual-data/developers/developers.csv
BESS cost curve (forward)
ModelledThe 2026 and 2030 points on the BESS $/kWh curve are modelled projections; historical points are sourced.
Refine in manual-data/policy/bess-cost-curve.csv
Update workflow
Edit a feed, rebuild, commit.
- Edit the relevant feed under
manual-data/<section>/…. - Run
npm run data:build(or:monthly/:quarterly) to regenerate the committed snapshots deterministically. - Commit the changed feed and snapshot. See
docs/ARCHITECTURE.md §9for the full recipe.
Glossary
Sector terms used across Solardash.
Glossary
18 terms