Skip to content
View nirajan550123's full-sized avatar

Block or report nirajan550123

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
nirajan550123/README.md

Nirajan Tripathi

Geospatial data scientist, Spatial machine learning, remote sensing, and reproducible analysis pipelines. M.S. Geography (GIS and remote sensing), Texas State University.

I build models on spatial data and care about whether they actually hold up: spatial cross-validation, reporting where a model stops working, and code someone else can run. My work spans energy siting, water resources, land-cover change, and urban environment, applied across the US and Nepal.

Open to GIS Analyst, GIS Developer, and geospatial data science roles. US, open to relocation and remote.

Featured work

Where Solar Gets Built: Utility-Scale Solar Siting in Texas A Random Forest model of utility-scale solar siting in Texas, with a North Carolina transfer test. Finding: distance to transmission infrastructure predicts siting roughly an order of magnitude better than solar irradiance. Spatial cross-validated ROC-AUC 0.92, with a full validation ladder and a cross-state transfer test that localizes exactly where the model stops generalizing. Python, scikit-learn, GeoPandas, Earth Engine, SHAP

Forecasting San Antonio Urban Growth to 2030 (Markov-CA) A Markov chain and Cellular Automata model that learns 2010 to 2020 land-cover transitions and forecasts where San Antonio grows by 2030, validated at Figure of Merit 0.85 against a persistence baseline. Python, ArcGIS Pro, NLCD

Rooftop Solar Potential: 1 m LiDAR vs a 30 m Open DSM The same rooftop-solar method run on 1 m airborne LiDAR and a 30 m open DSM, quantifying how input-data resolution systematically biases solar estimates. Python, ArcGIS Pro, USGS 3DEP, Copernicus GLO-30

Austin Urban Heat Island Story Map Summer land surface temperature across 65 Austin neighborhoods, measured in Earth Engine and joined to census income through a PostGIS spatial join. Canopy and heat correlate at r = -0.87. Earth Engine, PostGIS, R, Leaflet

More projects at nirajan550123.github.io.

Toolkit

Modeling and data: Python (scikit-learn, GeoPandas, Rasterio, pandas), R (tidyverse, sf), spatial cross-validation, SHAP Remote sensing: Google Earth Engine, Landsat and Sentinel, NDVI/LST, land-cover classification Spatial and database: PostgreSQL/PostGIS, spatial SQL, ArcGIS Pro, ArcPy, QGIS Delivery: Leaflet, GitHub Pages, reproducible project structure

Contact

Portfolio · LinkedIn · nirajantripathi5523@gmail.com

Pinned Loading

  1. solar-siting-tx solar-siting-tx Public

    ML analysis of utility-scale solar siting in Texas with a North Carolina transfer test. Grid access predicts siting ~9x better than irradiance. Random Forest, spatial CV, SHAP, AOA.

    Python

  2. lidar-rooftop-solar lidar-rooftop-solar Public

    Rooftop solar potential estimated with one method across two elevation-data resolutions: 1 m airborne LiDAR (Austin) vs 30 m open DSM (Kathmandu). Quantifies how coarse data biases the result.

    Python

  3. watershed-lulc-markov watershed-lulc-markov Public

    Markov-CA model forecasting urban land-cover change in the San Antonio watershed, validated at Figure of Merit 0.85

    Python

  4. austin-heat-storymap austin-heat-storymap Public

    Interactive scrollytelling map of Austin urban heat: summer land surface temperature vs tree canopy, impervious surface, and income across 65 neighborhoods. Earth Engine, PostGIS, R, Leaflet.

    HTML