Spaces:
Sleeping
Sleeping
| import gradio as gr | |
| from gradio_folium import Folium | |
| import folium | |
| import pandas as pd | |
| import numpy as np | |
| from geopy.distance import geodesic | |
| def calculate_distances(coords): | |
| distances = [] | |
| num_coords = len(coords) | |
| shortest_distances = [] | |
| for i in range(num_coords): | |
| shortest_distance = float('inf') | |
| closest_pair = None | |
| for j in range(num_coords): | |
| if i != j: | |
| distance = geodesic(coords[i], coords[j]).kilometers | |
| distances.append((distance, coords[i], coords[j])) | |
| if distance < shortest_distance: | |
| shortest_distance = distance | |
| closest_pair = (coords[i], coords[j]) | |
| if closest_pair: | |
| shortest_distances.append((shortest_distance, closest_pair[0], closest_pair[1])) | |
| return distances, shortest_distances | |
| def plot_on_map(coords_list, shortest_distance, longest_shortest_distance, avg_distance): | |
| all_coords = [coord for coords in coords_list for coord in coords] | |
| avg_lat = np.mean([coord[0] for coord in all_coords]) | |
| avg_lon = np.mean([coord[1] for coord in all_coords]) | |
| m = folium.Map(location=[avg_lat, avg_lon], zoom_start=10) | |
| colors = ['blue', 'red', 'green', 'purple', 'orange', 'darkblue', 'lightgray', 'darkgreen', 'cadetblue', 'pink'] | |
| for i, coords in enumerate(coords_list): | |
| group_marker_color = colors[i % len(colors)] | |
| for coord in coords: | |
| store_name = f"Store: {coord}" | |
| if len(coords) == 1: | |
| store_name = "Only 1 store" | |
| folium.Marker(location=coord, popup=store_name, icon=folium.Icon(color=group_marker_color)).add_to(m) | |
| if shortest_distance: | |
| folium.PolyLine(locations=[shortest_distance[1], shortest_distance[2]], color='green', weight=5, tooltip=f'Shortest Distance: {shortest_distance[0]:.2f} km').add_to(m) | |
| if longest_shortest_distance: | |
| folium.PolyLine(locations=[longest_shortest_distance[1], longest_shortest_distance[2]], color='red', weight=5, tooltip=f'Longest Among the Shortest Distances : {longest_shortest_distance[0]:.2f} km').add_to(m) | |
| return m | |
| def generate_map_with_stats(postcode_groups_input): | |
| df = pd.read_csv("speedmart99.csv") | |
| df['postcode'] = df['postcode'].astype(str).str.zfill(5) # Ensure postcodes are 5 digits long | |
| postcode_groups = postcode_groups_input.split(',') | |
| coords_list = [] | |
| for postcode_group_input in postcode_groups: | |
| key_length = len(postcode_group_input) | |
| # Filter rows where the postcode starts with the user input (considering the key length) | |
| group = df[df['postcode'].str[:key_length] == postcode_group_input] | |
| if not group.empty: | |
| coords = group['coords'].apply(lambda x: tuple(map(float, x.strip('()').split(',')))).tolist() | |
| coords_list.append(coords) | |
| # Combine all coordinates into a single list | |
| all_coords = [coord for coords in coords_list for coord in coords] | |
| if all_coords: | |
| distances, shortest_distances = calculate_distances(all_coords) | |
| if shortest_distances: | |
| shortest_distance = min(shortest_distances, key=lambda x: x[0]) | |
| longest_shortest_distance = max(shortest_distances, key=lambda x: x[0]) | |
| avg_distance = np.mean([dist[0] for dist in shortest_distances]) | |
| map_object = plot_on_map([all_coords], shortest_distance, longest_shortest_distance, avg_distance) | |
| statistics = [] | |
| for i, postcode_group in enumerate(postcode_groups): | |
| statistics.append(f"Postcode Group {i+1}: {postcode_group}") | |
| statistics.append(f"Shortest Distance (green): {shortest_distance[0]:.2f} km") | |
| statistics.append(f"Longest Among the Shortest Distances (red): {longest_shortest_distance[0]:.2f} km") | |
| statistics.append(f"Average of the Shortest Distances: {avg_distance:.2f} km") | |
| return map_object, "\n".join(statistics) | |
| else: | |
| return None, "No data found for the specified postcode groups." | |
| iface = gr.Interface( | |
| fn=generate_map_with_stats, | |
| inputs="text", | |
| outputs=[Folium(), "text"], | |
| title="99 Speedmart Stores Analyzer", | |
| description="This app allows you to <b>input <a href='https://en.wikipedia.org/wiki/Postal_codes_in_Malaysia#:~:text=Postal%20codes%20in%20Malaysia%20%2D%20Wikipedia,History'>postcode areas</a></b> and <b>visualize</b> the locations of <a href='https://www.99speedmart.com.my/About'>99 Speedmart</a> stores within those areas.\n \nYou can <b>analyze the <span style='color: red;'>longest</span>, <span style='color: green;'>shortest</span> and average distances</b> between stores. \nEnter the first few digits of the postcode (up to 5 digits) and you may combine multiple postcode areas by separating them by commas.\n\n\nYou may try: \n\n 811,813 for Tebrau(811xx) and Kempas(813xx) \n\n 4 for Klang(4xxxx) \n\n 83 for Batu Pahat(83xxx)", | |
| article="\n\n\n*Store location data was last updated in February 2024.", | |
| ) | |
| iface.launch(share=True) | |