背景：該公司是全球大型連鎖快餐企業(yè)，為了解其手機(jī)點(diǎn)餐門店分布情況，爬取門店詳細(xì)地址，實(shí)現(xiàn)數(shù)據(jù)分析及可視化?？梢暬枰鉀Q的思路:

• 1.需要把門店地址轉(zhuǎn)換成經(jīng)緯度；
• 2.需要根據(jù)得到的經(jīng)緯度提取門店詳細(xì)的省份、城市等信息；
• 3.構(gòu)造Geopandas所需數(shù)據(jù)格式；
• 4.合并各省市2018年GDP（億元）和常住人口（萬人）數(shù)據(jù)。

導(dǎo)入相關(guān)包

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.pylab as plb
from  matplotlib import cm
import seaborn as sns
plt.rc('font', family='SimHei', size=18)# 顯示中文標(biāo)簽
plt.rcParams['axes.unicode_minus'] = False
sns.set()
%matplotlib inline
import geopandas 
from shapely.geometry import LineString,Point
import re
import requests
import json
import hashlib
from wordcloud import WordCloud
from matplotlib.patches import Polygon
from mpl_toolkits.basemap import Basemap
from matplotlib.collections import PatchCollection
import warnings
warnings.filterwarnings('ignore')

數(shù)據(jù)處理

導(dǎo)入數(shù)據(jù)

data=pd.read_excel(r"D:\PycharmProjects\BK_store.xls")
data.head()

數(shù)據(jù)查看

data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1150 entries, 0 to 1149
Data columns (total 2 columns):
storename    1150 non-null object
address      1150 non-null object
dtypes: object(2)
memory usage: 18.0+ KB

data.dtypes

storename    object
address      object
dtype: object

data.isnull().sum()#查看缺失值

storename    0
address      0
dtype: int64

data.duplicated().sum()#查看重復(fù)值

0

自定義函數(shù)，根據(jù)地址查詢經(jīng)緯度

data.storename=data.storename.astype(str)
data.address=data.address.astype(str)

def get_locationdata(address):
    requests.adapters.DEFAULT_RETRIES =5
    requests.session().keep_alive = False
    url='http://restapi.amap.com/v3/geocode/geo?key=1fbbc548b828f33ccc64f4613ae52497&s=rsv3&city=35&address={}'.format(address)
    headers = {'User-Agent': 'Mozilla/5.0'}
    response = requests.get(url, headers=headers)
    if 'location' in response.text:
        for i in json.loads(response.text)['geocodes']:
            return i.get('location')
    else:
        pass

get_locationdata('昆山市虹祺路243號后街商業(yè)廣場C5幢樓一樓部分商鋪')

'120.920690,31.411805'

data['loncationdetail']=data.address.apply(get_locationdata)

data[data.loncationdetail.isnull()]#查看缺失值

data.drop([1018],inplace=True)#索引1018為測試門店數(shù)據(jù)，刪除該行數(shù)據(jù)
data.loc[95].address='徐州市'+data.loc[95].address
data.loc[332].address='昆明市'+data.loc[332].address
data.loc[648].address='常州市'+data.loc[648].address
data.loc[740].address='昆明市'+data.loc[740].address
data.loc[1033].address='鎮(zhèn)江市'+data.loc[1033].address
data.loc[1095].address='嘉興市'+data.loc[1095].address

data['loncationdetail']=data.address.apply(get_locationdata)#之前缺失的行再執(zhí)行一次，獲取經(jīng)緯度信息

data[data.loncationdetail.isnull()]#查看缺失值

data['lon']=data.loncationdetail.str.split(',').apply(lambda x:x[0]).astype(float)##經(jīng)度
data['lat']=data.loncationdetail.str.split(',').apply(lambda x:x[1]).astype(float)#緯度

data.head()

data.isnull().sum()#再次查看缺失值

storename          0
address            0
loncationdetail    0
city               0
province           0
district           0
township           0
lon                0
lat                0
dtype: int64

自定義函數(shù)，根據(jù)經(jīng)緯度查詢詳細(xì)地址

def get_data(location):
    url='https://restapi.amap.com/v3/geocode/regeo?key=1fbbc548b828f33ccc64f4613ae52497&s=rsv3&city=35&location={}'.format(location)
    requests.adapters.DEFAULT_RETRIES =5
    requests.session().keep_alive = False
    headers = {'User-Agent': 'Mozilla/5.0'}
    response = requests.get(url, headers=headers)
    address_dict={}
    if 'regeocode' in response.text:
        address_detail=re.search('"addressComponent":(.*),(.*),(.*),(.*)',response.text).group(1)
        address_dict['city']=json.loads(address_detail)['city']
        address_dict["province"]=json.loads(address_detail)["province"]
        address_dict["district"]=json.loads(address_detail)["district"]
        address_dict["township"]=json.loads(address_detail)["township"]
    else:
        pass
    return address_dict 

get_data('119.397392,32.380924')

{'city': '揚(yáng)州市', 'province': '江蘇省', 'district': '邗江區(qū)', 'township': '邗上街道'}

data['city']=data.loncationdetail.apply(get_data).apply(lambda x:x['city'])#提取城市

data['province']=data.loncationdetail.apply(get_data).apply(lambda x:x['province'])#提取省份

data['district']=data.loncationdetail.apply(get_data).apply(lambda x:x['district'])#提取轄區(qū)

data['township']=data.loncationdetail.apply(get_data).apply(lambda x:x['township'])#提取街道

data.head()

for i in ['city','province','district','township']:#查看含有空列表的列
        print(str(i),':',(data[i].apply(len)==0).sum())

city : 294
province : 0
district : 10
township : 7

data[data.province.str.contains('市')].index#查看含有city列為空列表的行索引

Int64Index([   0,    5,    7,    9,   11,   14,   20,   21,   28,   29,
            ...
            1123, 1130, 1131, 1134, 1135, 1136, 1137, 1138, 1146, 1147],
           dtype='int64', length=294)

for i in data[data.province.str.contains('市')].index:
    data.loc[i,'city']=data.loc[i,'province']

data.head()

可視化分析

轉(zhuǎn)換點(diǎn)數(shù)據(jù)

xy = [Point(xy) for xy in zip(data.lon,data.lat)]
geo_data=geopandas.GeoDataFrame(data,geometry=xy)
geo_data.head()

導(dǎo)入中國地圖shp

gdf = geopandas.read_file(r"D:\PycharmProjects\中國地圖shp格式\china_basic_map\bou2_4p.shp",encoding='gbk')
gdf.head()

各省市門店分布空間分析

gdf.plot(figsize=(15,15), alpha=0.5, edgecolor='white',color='gray')
plt.gca().xaxis.set_major_locator(plt.NullLocator()) #去掉x軸刻度
plt.gca().yaxis.set_major_locator(plt.NullLocator()) #去掉y軸刻度

output_41_0.png

geo_data.plot(ax=gdf.plot(figsize=(15,15),legend=True,edgecolor='white',color='gray'),color='yellow',markersize=8)
plt.rc('font', family='SimHei', size=20)# 顯示中文標(biāo)簽
plt.rcParams['axes.unicode_minus'] = False
plt.title('B快餐公司微信手機(jī)點(diǎn)餐門店城市分布圖',size=20)
plt.gca().xaxis.set_major_locator(plt.NullLocator()) #去掉x軸刻度
plt.gca().yaxis.set_major_locator(plt.NullLocator()) #去掉y軸刻度

output_42_0.png

geo_data['shop_counts']=geo_data.groupby('province')['geometry'].transform(lambda x: x.count())

geo_data.head()

geo_data.plot(ax=gdf.plot(figsize=(15,15),legend=True,edgecolor='white',color='gray'),column='shop_counts', cmap='rainbow',scheme='quantiles',legend=True,figsize=(8,10)
             ,marker='o', markersize=100) #按個數(shù)多少疊加底色
plt.gca().xaxis.set_major_locator(plt.NullLocator()) #去掉x軸刻度
plt.gca().yaxis.set_major_locator(plt.NullLocator()) #去年y軸刻度

output_45_0.png

各省市門店數(shù)統(tǒng)計(jì)

#各省市門店數(shù)
province_pc=geo_data.groupby('province',as_index=False)['shop_counts'].count().sort_values(by='shop_counts',ascending=False)
province_pc.head(10)

#對含有自治區(qū)的字符串換行處理
import textwrap
for i in province_pc[province_pc.province.str.contains('區(qū)')].index.to_list():
     province_pc.loc[i,'province']=textwrap.fill( province_pc.loc[i,'province'],province_pc.loc[i,'province'].find('自'))

province_pc[province_pc.province.str.contains('區(qū)')]

import squarify
labels = province_pc.apply(lambda x: str(x[0]) + "\n (" + str(x[1]) + ")", axis=1)
sizes = province_pc['shop_counts'].values.tolist()
colors = [plt.cm.Set1(i/float(len(labels))) for i in range(len(labels))]
plt.figure(figsize=(10,12),dpi= 80)
squarify.plot(sizes=sizes, label=labels, color=colors, alpha=.8)
plt.rc('font',size=10)
plt.title('各省市門店數(shù)樹形圖')
plt.axis('off')#去除上邊框和右邊框刻度
plt.tick_params(top='off',right='off')##去除上邊框和右邊框刻度

output_50_0.png

• 結(jié)論1：門店主要集中以下區(qū)域：江蘇省、上海市、浙江省、廣東省、北京市及山東省

各省市門店數(shù)標(biāo)準(zhǔn)化分析：Z-Score

x=province_pc.shop_counts
province_pc['province_z']= (x - x.mean())/x.std()#各省市門店數(shù)量的z分位數(shù)
province_pc['colors'] = ['red' if x < 0 else 'green' for x in province_pc['province_z']]
province_pc.sort_values('province_z', inplace=True)
province_pc.reset_index(inplace=True)
province_pc.head(10)

plt.figure(figsize=(16,20), dpi= 80)

plt.hlines(y=province_pc.index, xmin=0, xmax=province_pc.province_z, color=province_pc.colors, alpha=0.6, linewidth=15)

for x, y, tex in zip(province_pc.province_z, province_pc.index, province_pc.province_z):

    t = plt.text(x, y, round(tex,2), horizontalalignment='right' if x < 0 else 'left',

                 verticalalignment='center', fontdict={'color':'red' if x < 0 else 'green', 'size':15})

plt.gca().set(ylabel='省市', xlabel='門店數(shù)量')

plt.yticks(province_pc.index, province_pc.province, fontsize=15)

plt.title('各省市門店數(shù)量Z-Score', fontdict={'size':20})

plt.grid(linestyle='--', alpha=0.1)

output_54_0.png

#門店數(shù)Top10省市
province_pc.sort_values(by='shop_counts',ascending=False).head(10).province.to_list()

['江蘇省', '廣東省', '上海市', '北京市', '浙江省', '山東省', '遼寧省', '四川省', '福建省', '天津市']

• 結(jié)論2：Z-Score衡量各省市門店數(shù)相比于整體均值的偏差水平。其中門店數(shù)Top10的省市Z-Score>0，其他省市Z-Score<0。

各省市門店數(shù)與GDP、常住人口數(shù)相關(guān)性分析

#導(dǎo)入各省市2018年GDP（億元）和常住人口（萬人）數(shù)據(jù)
gp=pd.read_csv(r"D:\PycharmProjects\分省年度GDP及人口數(shù)據(jù).csv",encoding="gbk")
gp.head()

#將各省份門店數(shù)、2018年GDP、常住人口數(shù)合并
result=pd.merge(province_pc.loc[:,['province','shop_counts']],gp,how='left',on='province')
corr = result.corr()#相關(guān)系數(shù)
corr

cmap = sns.diverging_palette(220,10, as_cmap=True) 
sns.heatmap(corr, cmap=cmap,center=0.5, annot=True)
plt.xticks(fontsize=10)
plt.yticks(fontsize=10)
plt.title('各省市2018年GDP、常住人口數(shù)與門店數(shù)熱度圖', fontsize=15)

output_60_1.png

• 結(jié)論3：各市門店數(shù)與2018年GDP、常住人口數(shù)均正相關(guān)，其中與GDP相關(guān)性較高。

門店數(shù)Top14城市

#各城市門店數(shù)
city_pc=geo_data.groupby('city',as_index=False)['shop_counts'].count().sort_values(by='shop_counts',ascending=False)
city_pc

140 rows × 2 columns

y=city_pc.shop_counts
city_pc['city_z']= (y - y.mean())/y.std()#各省市門店數(shù)量的z分位數(shù)
city_pc.sort_values('city_z', inplace=True,ascending=False)
city_pc.reset_index(inplace=True)
del city_pc["index"]

#門店數(shù)Top14城市
city_pc[city_pc.city_z>=1]

input_colors =cm.rainbow(np.arange(len(city_pc.values))-2/len(city_pc.values)-2) 
plt.rc('font', family='SimHei', size=18)# 顯示中文標(biāo)簽
plt.rcParams['axes.unicode_minus'] = False
plt.figure(figsize=(10,10))
plt.pie(city_pc.shop_counts.head(14),  startangle=70, labels=city_pc.city.head(14),  
    autopct='%1.2f%%', pctdistance=0.8,  
    labeldistance=1.1, textprops={'fontsize':15,'color':'black'},colors=input_colors)  
plt.legend(bbox_to_anchor=(1.05, 0), loc=3, borderaxespad=0)
plt.title('門店數(shù)Top14城市',fontsize=20)

output_66_1.png