数据来源

DataV.GeoAtlas地理小工具系列 (aliyun.com)

代码

import requests
import json
import pandas as pd


def get_feature_list(url):
    try:
        r = requests.get(url)
        data = json.loads(r.text)
        return data['features']
    except Exception as e:
        print()


if __name__ == '__main__':


    url = 'https://geo.datav.aliyun.com/areas_v3/bound/100000_full.json'
    df = []

    pattern = '\d+'
    data = get_feature_list(url)

    for province in data:
        print(province)
        province_code = province['properties']['adcode']
        if province_code == '100000_JD':
            continue
        province_name = province['properties']['name']
        province_lon, province_lat = province['properties']['center']
        print(province_code, province_name)

        url = 'https://geo.datav.aliyun.com/areas_v3/bound/' + str(province_code) + '_full.json'
        province_data = get_feature_list(url)
        if province_data is None:
            df.append([province_name, province_code, province_lon, province_lat,
                       '', '', '', '',
                       '', '', '', ''])
            continue

        for city in province_data:
            city_code = city['properties']['adcode']
            city_name = city['properties']['name']
            city_lon, city_lat = city['properties']['center']
            print(city_code, city_name)

            url = 'https://geo.datav.aliyun.com/areas_v3/bound/' + str(city_code) + '_full.json'
            district_data = get_feature_list(url)

            if district_data is None:
                df.append([province_name, province_code, province_lon, province_lat
                              , city_name, city_code, city_lon, city_lat,
                           '', '', '', ''])
                continue

            for district in district_data:
                district_code = district['properties']['adcode']
                district_name = district['properties']['name']
                lon, lat = district['properties']['center']

                df.append([province_name, province_code, province_lon, province_lat
                              , city_name, city_code, city_lon, city_lat
                              , district_name, district_code, lon, lat])
    df = pd.DataFrame(df, columns=['省名称', '省代码', '省经度', '省维度',
                                   '市名称', '市代码', '市经度', '市维度',
                                   '区县名称', '区县代码', '区县经度', '区县维度'])
    df.to_excel('/Users/liuchaoyang/Desktop/data.xlsx', encoding='gbk', index=False)

    print(df)