Research

Research Interest

• Machine learning and Deep learning models to assess the risk of Biological Invasion

• Mathematical and Statistical modelling in ecological research

Masters Project

Phase 1 (Semester III) : Spatial Data Handling Using Python (py)

• • Handled different type of spatial data in Python (vector, raster etc.)

  • Fitted different variogram models to spatially referenced data.

  • Python libraries used- numpy, pandas, matplotlib, mpl_toolkits, scipy, skgstat, shapely, geopandas, Descartes, rasterstats, rasterio.

  • In spatial regression, we did variogram model fiting and derived equation of spatial prediction or kriging. And we have used Dirichlet Distribution on spatial data. We have derived Dirichlet distribution and represented the geometric behaviour of dirichlet Distribution by varying parameter values.

Masters Project

Phase 2 (Semester IV) : Species Distribution Modelling using Python – A case study using Mikania Micrantha Kunth

• • Developed predictive models in Python for computing the probability of occurrence of the invasive plant Mikania micrantha under current climate conditions and future climatic scenarios.

  • Spatially referenced occurrence data collected from www.gbif.org and climate data from www.worldclim.org in raster format.

  • Developed Python program to perform Machine learning methods for model building (Logistic and Generalized Additive Models (GAMs)).

  • Python libraries used in the project:

  • For geographic projection: geopandas, shapely, osmnx, Descartes, Fiona, georasters, rasterio.

    • Predictive Model building: sklearn, rpy2 (R package – mgcv, dismo)

    • Model visualization: matplotlib, seaborn.

    • Part of the analysis was carried out by calling R from Python using rpy2 library.