5lines_kmeans.py

def scratch_k_means(k, all_data):
    
    
    centroids= [(random.randint(min(x_data), max(x_data)), random.randint(min(y_data), max(y_data))) for i in range(k)]
    for i in range(100):
        distances = np.concatenate([np.linalg.norm(all_data - centroid, axis=1).reshape(-1, 1) for centroid in centroids],axis=1)
        labels = np.argmin(distances, axis=1)
        centroids =pd.DataFrame(np.concatenate([all_data, labels.reshape(-1, 1)], axis=1)).groupby(2).mean().values
     
    
    
    return centroids, labels