fio-l2-capetown-landuse-dissolve

Rerun

import geopandas as gpd

def inspect_data(file_path):
    try:
        gdf = gpd.read_file(file_path)
        print(f"Total rows: {len(gdf)}")
        print(f"Columns: {gdf.columns.tolist()}")
        print(f"CRS: {gdf.crs}")
        print("First 5 rows:")
        print(gdf.head())
        print("\nValue counts for 'landuse' (or similar):")
        # Look for likely columns for landuse
        potential_cols = ['landuse', 'class', 'type', 'landuse_class']
        for col in potential_cols:
            if col in gdf.columns:
                print(f"--- {col} ---")
                print(gdf[col].value_counts())
    except Exception as e:
        print(f"Error: {e}")

if __name__ == "__main__":
    inspect_data('capetown_landuse.fgb')

import geopandas as gpd

def solve():
    input_file = 'capetown_landuse.fgb'
    output_file = 'landuse_dissolved.geoparquet'
    
    # Load the data
    print(f"Reading {input_file}...")
    gdf = gpd.read_file(input_file)
    
    # Ensure we have the right column name
    # Based on inspection, the column is 'class'
    landuse_col = 'class'
    
    print(f"Dissolving by {landuse_col}...")
    
    # We need to:
    # 1. Group by 'class'
    # 2. Dissolve geometries into single MultiPolygons
    # 3. Calculate total area in m2 (CRS is EPSG:32734, which is metric)
    # 4. Count the number of source parcels
    
    # First, calculate individual parcel areas to be safe, 
    # though we will sum them after dissolution or calculate from the resulting geometry.
    # Let's calculate the area of the dissolved geometries to be most accurate to the final shape.
    
    # Aggregation
    # Groupby 'class' and aggregate:
    # - geometry: union (dissolve)
    # - id: count (parcel_count)
    
    # Dissolve handles the geometry union.
    # To get the count, we can use a groupby aggregation or perform it separately.
    
    # Let's do it in a way that is clean:
    stats = gdf.groupby(landuse_col).agg(
        parcel_count=('id', 'count'),
        geometry=('geometry', 'dissolve') # Note: dissolve is a method on GDF, not a string in agg for all versions
    )
    # Wait, gdf.dissolve(by='class') is the standard geopandas way.
    
    dissolved = gdf.dissolve(by=landuse_col, aggfunc={'id': 'count'})
    
    # Rename columns
    dissolved = dissolved.rename(columns={'id': 'parcel_count'})
    
    # The index is now 'class'. Let's reset it to make 'class' a column.
    dissolved = dissolved.reset_index()
    
    # Calculate area in m2
    # Since CRS is EPSG:32734 (UTM zone 34S), units are meters.
    print("Calculating areas...")
    dissolved['area_m2'] = dissolved.geometry.area
    
    # Reorder columns as requested: class, area_m2, parcel_count
    # (and keep geometry for the GeoParquet format)
    result = dissolved[['class', 'area_m2', 'parcel_count', 'geometry']]
    
    # Save to GeoParquet
    print(f"Writing to {output_file}...")
    result.to_parquet(output_file)
    print("Done!")

if __name__ == "__main__":
    solve()

import geopandas as gpd

def solve():
    input_file = 'capetown_landuse.fgb'
    output_file = 'landuse_dissolved.geoparquet'
    
    # Load the data
    print(f"Reading {input_file}...")
    gdf = gpd.read_file(input_file)
    
    # Ensure we have the right column name
    # Based on inspection, the column is 'class'
    landuse_col = 'class'
    
    print(f"Dissolving by {landuse_col}...")
    
    # We need to:
    # 1. Group by 'class'
    # 2. Dissolve geometries into single MultiPolygons
    # 3. Calculate total area in m2 (CRS is EPSG:32734, which is metric)
    # 4. Count the number of source parcels
    
    # Dissolve handles the geometry union and aggregation.
    # Using id to count parcels.
    dissolved = gdf.dissolve(by=landuse_col, aggfunc={'id': 'count'})
    
    # Rename columns
    dissolved = dissolved.rename(columns={'id': 'parcel_count'})
    
    # The index is now 'class'. Let's reset it to make 'class' a column.
    dissolved = dissolved.reset_index()
    
    # Calculate area in m2
    # Since CRS is EPSG:32734 (UTM zone 34S), units are meters.
    print("Calculating areas...")
    dissolved['area_m2'] = dissolved.geometry.area
    
    # Reorder columns as requested: class, area_m2, parcel_count
    # (and keep geometry for the GeoParquet format)
    result = dissolved[['class', 'area_m2', 'parcel_count', 'geometry']]
    
    # Save to GeoParquet
    print(f"Writing to {output_file}...")
    result.to_parquet(output_file)
    print("Done!")

if __name__ == "__main__":
    solve()