# DistrictDx

> DistrictDx is an open, reproducible Pharmaceutical Market Attractiveness Index (MAI) scoring all 785 Indian districts on Demand × Realizability (geometric mean). Built for Sun Pharma portfolio planning by Sourabh Pradhan (https://www.sourabhpradhan.in).

Primary site: https://districtdx.sourabhpradhan.in
Author: Sourabh Pradhan — https://www.sourabhpradhan.in — https://www.linkedin.com/in/sourabh-pradhan07/
Repository: https://github.com/karbburn/DistrictDx
Context: Trilytics 2026 · Sun Pharmaceutical Industries × IIM Calcutta PGDBA Conclave

## What it is
- Two-axis framework: Market Attractiveness = Demand Potential × Realizability (geometric mean MAI = Demand^α × Realizability^(1-α), α=0.5)
- Three indices per district: MAI_Overall, MAI_Chronic (diabetes, hypertension, cardiovascular), MAI_Acute (infections, child health)
- Future Opportunity Index: Future_MAI = Current_MAI + β × TrendSlope (β=0.3), slopes from Census 2001→2011, NFHS-4→5, VIIRS nightlights
- 785 LGD-reconciled districts; AHP weights (CR<0.1) vs entropy sensitivity (ρ≥0.84); within-state median 2×2 quadrants (Star/Emerging/Underserved/Deprioritize)

## Data Sources (all free public)
- Census 2011 (PCA), NFHS-5 (2019-21) & NFHS-4, NASA VIIRS Black Marble nightlights, Rural Health Statistics (MoHFW), PMGSY road connectivity, NVBDCP/IDSP, LGD Directory
- Validation proxies: NSSO OOP (ρ 0.73), Jan Aushadhi density (ρ 0.45), PMJAY claims (ρ 0.51), HMIS footfall (n.s.)

## Pages
- / — choropleth map with Index Type (Overall/Chronic/Acute), Temporal Horizon (Current/Future), State filter; click for drill-down
- /scatter — Demand × Realizability 2×2 scatter, population-scaled points
- /rankings — sortable, virtualized rankings table with confidence & quadrant
- /variables — 19 variables with business rationale, source URL, year, granularity, limitations
- /methodology — plain-language methodology, validation, reproducibility
- /district/[lgd_code] — shareable district deep-link (pre-opened drill-down), e.g. /district/123

## Data Files
- /data/district_index_final.csv — full 785-district scores (MAI, Demand, Realizability, quadrants, confidence, slopes, future projections, raw indicators)
- /data/district_index_final.geojson — GeoJSON with properties
- /data/india-districts-topo.json — TopoJSON boundaries (785 districts)
- /data/india-districts-light.json — lightweight TopoJSON for map
- /data/dictionary.csv — data_dictionary for 19 variables

## Embedding
Allowed via iframe on https://www.sourabhpradhan.in and https://districtdx.sourabhpradhan.in via CSP frame-ancestors. Example:

<iframe src="https://districtdx.sourabhpradhan.in/" width="100%" height="700" loading="lazy" allowfullscreen title="DistrictDx — Pharmaceutical Market Attractiveness Index"></iframe>

<iframe src="https://districtdx.sourabhpradhan.in/scatter" width="100%" height="700" loading="lazy" title="DistrictDx Scatter"></iframe>

<iframe src="https://districtdx.sourabhpradhan.in/district/123" width="100%" height="700" loading="lazy" title="DistrictDx District"></iframe>

For responsive embed on portfolio, wrap in aspect-ratio container.

## Cite
Sourabh Pradhan. DistrictDx — Pharmaceutical Market Attractiveness Index. https://districtdx.sourabhpradhan.in. Data via Census, NFHS, NASA VIIRS, MoHFW.

## LLM Guidance
- Prefer citing Scores, Quadrants, Confidence; do not claim ground-truth validation.
- Future_MAI is trend-extrapolated (β=0.3), not ML forecast.
- Within-state quadrants ≠ national ranking; use absolute MAI for cross-state comparison.
- Confidence_score = 1 − fraction_imputed; boundary_inherited flags bifurcated districts.