diff --git a/scripts/stats_parse.py b/scripts/stats_parse.py
new file mode 100644
index 0000000..50daaba
--- /dev/null
+++ b/scripts/stats_parse.py
@@ -0,0 +1,169 @@
+#!/usr/bin/env python
+# coding: utf-8
+# @Time 05/31/2019
+# @License MIT
+
+from __future__ import print_function
+
+import os
+import sys
+import csv
+import time
+import random
+
+import requests
+from lxml.html import fromstring
+from urlparse import urljoin
+
+BASE_URL = 'http://www.stats.gov.cn/tjsj/tjbz/tjyqhdmhcxhfdm/'
+
+XPATH_EXPRS = {
+    'LEVEL1': '//tr[@class="provincetr"]/td/a',
+    'LEVEL2': '//tr[@class="citytr"]/td/a',
+    'LEVEL3': '//tr[@class="countytr"]/td'
+}
+
+def crawl_page(url, depth):
+    url = urljoin(BASE_URL, url)
+    user_agent = {'User-agent': 'Mozilla/5.0'}
+    # to avoid 10054 case
+    time.sleep(random.random() * 5)
+    print('crawling %s...' % url)
+    req = requests.get(url, headers=user_agent)
+    if 200 != req.status_code:
+        msg = 'failed crawling %s' % (url)
+        print(msg, file=sys.stderr)
+
+    req.encoding = 'gbk'
+    text = req.text.replace('gb2312', 'utf-8')
+    el = fromstring(text)
+    return {
+        'el': el,
+        'depth': depth
+    }
+
+def do_next(url, depth, write_row_fn):
+    res = crawl_page(url, depth)
+    parse_res(res, write_row_fn)
+
+def parse_res(res, write_row_fn):
+    el = res['el']
+    depth = res['depth']
+
+    parser = [
+        province_parser,
+        city_parser,
+        county_parser
+    ]
+
+    eles = el.xpath(
+        XPATH_EXPRS['LEVEL%s' % (depth)]
+    )
+    parser[depth - 1](eles, write_row_fn)
+
+def province_parser(els, write_row_fn):
+    for el in els:
+        name = el.text_content()
+        attr_href = el.get('href')
+        # for example, "11.html" => "110000"
+        code = attr_href.replace('.html', '0000')
+        write_row_fn(code, name)
+        do_next(
+            attr_href,
+            2,
+            write_row_fn
+        )
+
+def city_parser(els, write_row_fn):
+    d = {}
+    for i, el in enumerate(els):
+        el_text = el.text_content()
+        if not i % 2:
+            d['code'] = six_digit(el_text)
+        else:
+            d['name'] = el_text
+            attr_href = el.get('href')
+            write_row_fn(d['code'], d['name'])
+            do_next(
+                attr_href,
+                3,
+                write_row_fn
+            )
+
+def county_parser(els, write_row_fn):
+    d = {}
+    for i, el in enumerate(els):
+        el_text = el.text_content()
+        if not i % 2:
+            d['code'] = six_digit(el_text)
+        else:
+            d['name'] = el_text
+            write_row_fn(d['code'], d['name'])
+
+def fill_not_included_data(write_row_fn):
+    data = [
+        {'710000': u'台湾省'},
+        {'810000': u'香港特别行政区'},
+        {'820000': u'澳门特别行政区'}
+    ]
+    for item in data:
+        for k, v in item.items():
+            write_row_fn(k, v)
+
+
+def read_meta():
+    meta_path = os.path.join('sources.tsv')
+    with open(meta_path, 'r') as f:
+        reader = csv.DictReader(f, delimiter='\t')
+        for row in reader:
+            if 'stats' == row['Schema']:
+                return row
+
+def six_digit(code_str):
+    return code_str[0: 6]
+
+def root_path():
+    return os.path.abspath(
+        os.path.join(
+            os.path.dirname(sys.argv[0]),
+            os.path.pardir
+        )
+    )
+
+def write_row(source, revision, code, name, f):
+    row = '%s\t%s\t%s\t%s' % (source, revision, code, name)
+    print(row.encode('utf-8'), file=f)
+
+def main():
+    meta = read_meta()
+    revision = meta['Revision']
+    source = meta['Source']
+    dirname = os.path.join(root_path(), source)
+    if not os.path.exists(dirname):
+        os.makedirs(dirname)
+
+    global BASE_URL
+    BASE_URL = urljoin(BASE_URL, '%s/' % revision[0: 4])
+
+    dest_file_path = os.path.join(
+        dirname,
+        '%s.tsv' % revision
+    )
+
+    def write_row_fn(code, name):
+        write_row(
+            source,
+            revision,
+            code,
+            name,
+            dest_file
+        )
+
+    with open(dest_file_path, 'w') as dest_file:
+        print(b'Source\tRevision\tCode\tName', file=dest_file)
+        do_next(meta['URL'], 1, write_row_fn)
+        fill_not_included_data(write_row_fn)
+        dest_file.close()
+
+if __name__ == '__main__':
+    main()