中国高等学校数据分析及可视化

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib as mpl
import seaborn as sn
from pylab import *
mpl.rcParams['font.sans-serif']=['SimHei']
from pyecharts.charts import Map
from pyecharts import options as opts

#一、读取数据：
data = pd.read_excel("全国高等学校名单.xls")
data=data.drop(['备注'],axis=1)

#统计在京本科高校：

#数据处理：
data_beijing=data[(data['所在地']=='北京市')&(data['办学层次']=='本科')]      #多条件筛选
data_beijing_department=data_beijing.groupby('主管部门').size().reset_index()

#数据可视化
fig,axes=plt.subplots()
fig.set_size_inches(20,12)
sn.barplot(data=data_beijing_department,x="主管部门",y=0)
axes.set(xlabel='主管部门',ylabel='高校数量',title='在京高校所属部门分布图')
plt.savefig("在京高校所属部门分布图.png")
plt.show()


#统计全国各地本科高校数量
#数据处理
data_university=data[data['办学层次']=='本科']
data_university=data_university.groupby('所在地（省）').size().reset_index()

#数据可视化1
provinces=list(data_university['所在地（省）'])
values=list(data_university.iloc[:,1])
map = Map(init_opts=opts.InitOpts(width="1900px", height="900px", bg_color="#ADD8E6",
	       page_title="中国高校分布图",theme="white"))

map.add("数量：",[list(z) for z in zip(provinces,values)],is_map_symbol_show=False,
        maptype="china",label_opts=opts.LabelOpts(is_show=False),
        itemstyle_opts=opts.ItemStyleOpts(color="rgb(49,60,72)"))

map.set_global_opts(title_opts = opts.TitleOpts(title='中国高校分布图'),
	                legend_opts=opts.LegendOpts(is_show=False),
                    visualmap_opts = opts.VisualMapOpts(max_=100000))
map.render(path='中国高校分布图1.html')

#数据可视化2
map = Map(init_opts=opts.InitOpts(width="1900px", height="900px", bg_color="#ADD8E6",
	       page_title="中国高校分布图",theme="white"))
map.set_global_opts(
    title_opts=opts.TitleOpts(title="中国高校分布图"),
    visualmap_opts=opts.VisualMapOpts(max_=100000, is_piecewise=True,
                                      pieces=[
                                        {"max": 100, "min": 61, "label": ">60", "color": "#333399"},
                                        {"max": 60, "min": 51, "label": "50-60", "color": "#3333CC"},
                                        {"max": 50, "min": 41, "label": "40-50", "color": "#3366CC"},
                                        {"max": 40, "min": 31, "label": "30-40", "color": "#3366FF"},
                                        {"max": 30, "min": 21, "label": "20-30", "color": "#3399FF"},
                                        {"max": 20, "min": 11, "label": "10-20", "color": "#33CCFF"},
                                        {"max": 10, "min": 0, "label": "0-10", "color": "#33FFFF"},
                                        ], ) 
    )
map.add("中国高校分布图", data_pair=[list(z) for z in zip(provinces,values)], maptype="china", label_opts=opts.LabelOpts(is_show=True),is_roam=True)
map.render('中国高校分布图2.html')