這篇文章給出了如何繪制中國(guó)人口密度圖，但是運(yùn)行存在一些問(wèn)題，我在一些地方進(jìn)行了修改。

本人使用的IDE是anaconda，因此事先在anaconda prompt 中安裝Basemap包

conda install Basemap

新建文檔，導(dǎo)入需要的包

import matplotlib.pyplot as pltfrom mpl_toolkits.basemap import Basemapfrom matplotlib.patches import Polygonfrom matplotlib.colors import rgb2heximport numpy as npimport pandas as pd

Basemap中不包括中國(guó)省界，需要在下面網(wǎng)站下載中國(guó)省界，點(diǎn)擊Shapefile下載。

生成中國(guó)大陸省界圖片。

plt.figure(figsize=(16,8))m = Basemap( llcrnrlon=77, llcrnrlat=14, urcrnrlon=140, urcrnrlat=51, projection='lcc', lat_1=33, lat_2=45, lon_0=100)m.drawcountries(linewidth=1.5)m.drawcoastlines() m.readshapefile('gadm36_CHN_shp/gadm36_CHN_1', 'states', drawbounds=True)

去國(guó)家統(tǒng)計(jì)局網(wǎng)站下載人口各省，只需保留地區(qū)和總?cè)丝诩纯桑４鏋閏sv格式并改名為pop.csv。

讀取數(shù)據(jù)，儲(chǔ)存為dataframe格式，刪去地名之中的空格，并設(shè)置地名為dataframe的index。

df = pd.read_csv('pop.csv')new_index_list = []for i in df["地區(qū)"]: i = i.replace(" ","") new_index_list.append(i)new_index = {"region": new_index_list}new_index = pd.DataFrame(new_index)df = pd.concat([df,new_index], axis=1)df = df.drop(["地區(qū)"], axis=1)df.set_index("region", inplace=True)

將Basemap中的地區(qū)與我們下載的csv中的人口數(shù)據(jù)對(duì)應(yīng)起來(lái)，建立字典。注意，Basemap中的地名與csv文件中的地名并不完全一樣，需要進(jìn)行一些處理。

provinces = m.states_infostatenames=[]colors = {}cmap = plt.cm.YlOrRdvmax = 100000000vmin = 3000000 for each_province in provinces: province_name = each_province['NL_NAME_1'] p = province_name.split('|') if len(p) > 1:  s = p[1] else:  s = p[0] s = s[:2] if s == '黑龍':  s = '黑龍江' if s == '內(nèi)蒙':  s = '內(nèi)蒙古' statenames.append(s) pop = df['人口數(shù)'][s] colors[s] = cmap(np.sqrt((pop - vmin) / (vmax - vmin)))[:3]

最后畫(huà)出圖片即可

ax = plt.gca()for nshape, seg in enumerate(m.states): color = rgb2hex(colors[statenames[nshape]]) poly = Polygon(seg, facecolor=color, edgecolor=color) ax.add_patch(poly) plt.show()

完整代碼如下

# -*- coding: utf-8 -*- import matplotlib.pyplot as pltfrom mpl_toolkits.basemap import Basemapfrom matplotlib.patches import Polygonfrom matplotlib.colors import rgb2heximport numpy as npimport pandas as pd plt.figure(figsize=(16,8))m = Basemap( llcrnrlon=77, llcrnrlat=14, urcrnrlon=140, urcrnrlat=51, projection='lcc', lat_1=33, lat_2=45, lon_0=100)m.drawcountries(linewidth=1.5)m.drawcoastlines() m.readshapefile('gadm36_CHN_shp/gadm36_CHN_1', 'states', drawbounds=True) df = pd.read_csv('pop.csv')new_index_list = []for i in df["地區(qū)"]: i = i.replace(" ","") new_index_list.append(i)new_index = {"region": new_index_list}new_index = pd.DataFrame(new_index)df = pd.concat([df,new_index], axis=1)df = df.drop(["地區(qū)"], axis=1)df.set_index("region", inplace=True) provinces = m.states_infostatenames=[]colors = {}cmap = plt.cm.YlOrRdvmax = 100000000vmin = 3000000 for each_province in provinces: province_name = each_province['NL_NAME_1'] p = province_name.split('|') if len(p) > 1:  s = p[1] else:  s = p[0] s = s[:2] if s == '黑龍':  s = '黑龍江' if s == '內(nèi)蒙':  s = '內(nèi)蒙古' statenames.append(s) pop = df['人口數(shù)'][s] colors[s] = cmap(np.sqrt((pop - vmin) / (vmax - vmin)))[:3] ax = plt.gca()for nshape, seg in enumerate(m.states): color = rgb2hex(colors[statenames[nshape]]) poly = Polygon(seg, facecolor=color, edgecolor=color) ax.add_patch(poly) plt.show()