I will do remote sensing machine learning with random forest, xgboost and shap
GIS and Remote Sensing Expert for InSAR and Google Earth Engine
About this Gig
Machine learning for satellite and environmental data, done the way reviewers expect. I am a remote sensing PhD researcher at Tianjin University (MS CUG Wuhan) and build ML models for land cover, water, heat and hazard mapping. Example in my gallery: XGBoost + SHAP explaining urban land surface temperature (R2 0.75 on a spatially held-out test zone).
WHAT I DO: land cover / crop / forest classification (Random Forest, SVM, XGBoost, LightGBM); regression and prediction (water quality, chlorophyll-a, LST, air pollution, biomass, soil moisture, yield); explainable AI with SHAP and feature importance; spatial cross-validation (no data leakage); uncertainty maps and prediction intervals; accuracy assessment (OA, kappa, R2, RMSE).
DATA: Sentinel-1/2, Landsat, MODIS, VIIRS, GEDI, ERA5, in-situ and field data. Tools: Python (scikit-learn, XGBoost, SHAP, PyTorch), R, Google Earth Engine.
You get clean, commented code, maps and plots ready for a paper or report, and a short methods write-up. Message me with your data and goal first and I will suggest the right model.
Programming language:
Python
•
R
•
MATLAB
•
Colab
Frameworks:
Scikit-learn
•
Keras
•
PyTorch
•
Panda
Tools:
Jupyter Notebook
•
Excel
•
Colab
•
RStudio
Other Data Science & ML Services I Offer
FAQ
Can I bring my own data?
Yes. Send your CSV, rasters or shapefiles. If you do not have data, I can build the dataset from free satellite archives.
Will the results be publishable?
I follow good practice for papers: spatial cross-validation, independent test sets, uncertainty and clear methods text. Many of my figures are from peer-reviewed work.

