Bhargav Chintam

Projects

Click a title for the full write-up (method, formulas, real data sources). The interactive maps open directly in the browser. Numbers shown are the actual output of each project's demo run (synthetic scene, seed 42).

Urban planning

Urban Heat Island Analysis

Landsat 8/9 thermal band to land surface temperature, UHI intensity classes, land-cover correlation and a Mitigation Priority Index that tells planners where trees and cool roofs help most.

Mean LST 43.2 C, max 55.4 C. LST vs impervious surface r = 0.82. 378 ha flagged high priority.

Single-channel LSTNDVI emissivityZonal statsWeighted index

Interactive map Read more Dashboard code

Hydrology and emergency planning

Multi-Criteria Flood Risk Assessment

Priority-flood DEM conditioning, D8 flow and TWI, a seven-criteria AHP hazard model, socio-economic vulnerability and a composite risk index with 10 to 500 year scenarios.

7.0 % of the area High or Very High risk, 10,202 residents exposed, 3 of 25 critical facilities at risk. AHP CR = 0.011.

Priority floodD8 / TWIAHPRisk matrix

Interactive map Read more Dashboard code

Precision agriculture

Multi-Temporal Crop Health Monitoring

A full Sentinel-2 season: cloud masking, five vegetation indices, double-logistic phenology fitting, stress anomaly detection and per-field health reports for growers and agronomists.

14 scenes, peak NDVI 0.81 at day 207, 2.7 % stressed area in 7 clusters, all 4 nitrogen-stressed fields caught by NDRE.

Sentinel-2 L2APhenologyIsolation ForestZonal stats

Interactive map Read more Dashboard code

Public health and network analysis

Healthcare Accessibility and Site Selection

Road-network travel times, drive-time isochrones, two-step floating catchment accessibility scores, underserved block detection and a greedy maximal-coverage model for placing new facilities.

89.3 % of residents within 15 min of a hospital, 100 % after two proposed sites. 101,714 underserved people located.

DijkstraIsochrones2SFCAFacility location

Interactive map Read more Dashboard code

Forestry and environment

Forest Change Detection

Two-date Sentinel-2 Random Forest classification, post-classification change combined with dNBR, minimum-mapping-unit filtering, loss polygons and a full accuracy assessment.

44.8 ha of forest loss (3.2 %, 0.63 % per year) in 29 patches, 87 % within 500 m of roads. OA 0.98, kappa 0.97.

Random ForestdNBRChange detectionKappa

Interactive map Read more Dashboard code

Energy and infrastructure

Utility-Scale Solar Site Suitability

Terrain derivatives, standardized criterion curves, AHP weights, hard constraint masks and a weighted overlay that yields ranked candidate parcels with MW and GWh estimates.

10 candidate sites, 1,911 MW developable, best site 549 MW. 35 % of the area excluded by constraints. AHP CR = 0.013.

Slope / aspectAHPConstraintsSite ranking

Interactive map Read more Dashboard code

Run any project yourself

No accounts, no downloads. Each demo builds a realistic georeferenced synthetic scene, runs the real pipeline code and writes rasters, vectors, figures and a web map.

git clone https://github.com/bhargavchintam/qgis-geospatial-projects.git
cd qgis-geospatial-projects/flood-risk-assessment   # or any other project
pip install -r requirements.txt
python demo.py              # 2 to 10 seconds
streamlit run app.py        # interactive dashboard
python -m pytest tests -q   # tests

To use real data, point each project's config.yaml at your own rasters and vectors. Every README lists the public sources (USGS, Copernicus, NOAA, NLCD, Census, OpenStreetMap, NREL, HIFLD).

Skills shown across the portfolio

Grouped by what a GIS or geospatial data role usually asks for.