EmmanuelBotchway/FutureLandCoverPredictionCA-Markov

Forecasting Urban Landcover Dynamics in Obuasi Municipality, Ghana: # CA-Markov Obuasi Landcover Forecasting Project This package contains separate Jupyter notebooks and reusable Python functions for the CA-Markov landcover forecasting study for Obuasi Municipality, Ghana. ## Project purpose The workflow uses Hybrid GNN-ViT multimodal classified landcover maps from 2018-2025 to forecast landcover dynamics from 2026-2030 using CA-Markov modelling. ## Folder structure ```text ca_markov_obuasi_project/ ├── Data/ │ ├── input_rasters/ │ └── boundary/ ├── notebooks/ │ ├── 00_main_ca_markov_pipeline.ipynb │ ├── 01_project_setup.ipynb │ ├── 02_raster_loading_preprocessing.ipynb │ ├── 03_historical_landcover_dynamics.ipynb │ ├── 04_transition_matrix_analysis.ipynb │ ├── 05_model_validation_backcasting.ipynb │ ├── 06_future_landcover_forecasting.ipynb │ └── 07_spatial_change_gain_loss_analysis.ipynb ├── src/ │ └── ca_markov_utils.py ├── outputs/ │ ├── plots/ │ ├── tables/ │ ├── predicted_maps/ │ ├── validation/ │ └── change_maps/ └── docs/ ``` ## Required input files Place the classified GeoTIFF maps in: ```text Data/input_rasters/ ``` Use these filenames: ```text landcover_2018.tif landcover_2019.tif landcover_2020.tif landcover_2021.tif landcover_2022.tif landcover_2023.tif landcover_2024.tif landcover_2025.tif ``` ## Landcover class codes | Code | Class | |---:|---| | 1 | Water | | 2 | Trees | | 3 | Crops | | 4 | Built-up Areas | | 5 | Bare Ground | ## How to run Open: ```text notebooks/00_main_ca_markov_pipeline.ipynb ``` Run all cells. The main notebook calls the other notebooks in sequence. ## Required Python packages Install the following packages if needed: ```bash pip install numpy pandas matplotlib rasterio geopandas openpyxl ``` If `geopandas` is not needed for your current workflow, it can be omitted. `rasterio` is required for reading and writing GeoTIFF files. ## Main outputs The workflow produces: - Historical landcover statistics table - Historical trend graph - Average tr …