clustering.py

import pandas as pd
import geopandas as gpd
import numpy as np
from sklearn.cluster import DBSCAN

def perform_clustering(input_csv, output_csv):
    df = pd.read_csv(input_csv)
    df = df.dropna(subset=['LATITUDE', 'LONGITUDE'])

    gdf = gpd.GeoDataFrame(
        df, 
        geometry=gpd.points_from_xy(df.LONGITUDE, df.LATITUDE), 
        crs='EPSG:4326'
    )
    gdf = gdf.to_crs(epsg=3857)
    coords = np.vstack([gdf.geometry.x, gdf.geometry.y]).T

    db = DBSCAN(eps=50, min_samples=50, metric='euclidean')
    labels = db.fit_predict(coords)
    gdf['cluster'] = labels

    gdf.drop(columns='geometry').to_csv(output_csv, index=False)
    print(f"Clustered data saved to {output_csv}")

if __name__ == "__main__":
    input_csv = "data/dataset.csv"
    output_csv = "data/clustered_dataset.csv"
    perform_clustering(input_csv, output_csv)