data visualization--US Election 2020

数据准备阶段

这里主要用到的是这个总统候选人城镇（./data/president_county_candidate.csv）的这个表来做数据展示

数据操作

数据的导入

path_president_candidate = './data/president_county_candidate.csv'
usa_election =pd.read_csv(path_president_candidate)
usa_election

从这个数据我们可以看出美国的选举包括51个州（state）和n多个城市（county）以及候选人（candition）的名字（Joe Biden、Donald Trump、Jo Jorgensen、Howie Hawkins），以及所代表的党派（party），和总共获得的票数（total_votes）,以及是否获胜（won）

usa_election.info()

定义一个函数对输入的州和城镇的名字，画出相印的候选人的得票情况

def county_result(county,state):
    data = usa_election[(usa_election["county"] ==county)&(usa_election["state"] ==state)]
    fig = px.bar(data,x="candidate",y="total_votes")
    fig.update_layout(
        title={
        "text":"Election result in {country}({state})",
        "y":0.95,
        "x":0.5
    },
        xaxis_title="candidate",
        yaxis_title="total_votes",
        
    )
    fig.show()

county_result("Kent County","Delaware")

对这样子的数据我们要把Donald Trump和Joe Biden所代表的dem和rep单独的拎出来进行数据分析

dem的数据操作

#把共和党的选票按州的形式聚集起来
dem_election =usa_election[usa_election['party']=="DEM"]
dem_election = dem_election.groupby('state').sum()
dem_election.rename(columns ={"total_votes":"dem_votes","won":"dem_won_counties"},inplace = True)
dem_election.head(5)

rep的数据操作

rep_election =usa_election[usa_election['party']=="REP"]
rep_election = rep_election.groupby('state').sum()
rep_election.rename(columns ={"total_votes":"rep_votes","won":"rep_won_counties"},inplace = True)
rep_election.head(5)

这样子我们就得到了我们需要的数据集即DEM/REP在各个州的赢了多少个城镇以及得到的具体票数

画出DEM/REP的总共的得票总量以及占比

def most_voted_candidate():
    total_votes_dem = dem_election["dem_votes"].sum()
    total_votes_rep = rep_election["rep_votes"].sum()
    total_vote = usa_election.total_votes.sum()
    print(f"total votes for jeo briden(DEM):{total_votes_dem}")
    print(f"total votes for jeo Donald Trump(REP):{total_votes_rep}")
    fig =px.bar(x=['DEM',"REP"],y=[total_votes_dem,total_votes_rep])
    
    fig.update_layout(
        title = {
        "text":"total votes for dem and rep",
        "y":0.95,
        "x":0.5
    },
        xaxis_title = "parties",
        yaxis_title = "total votes"
    )
    fig.show()
    
    fig = px.bar (x=['DEM',"REP"] , y = [(total_votes_dem/total_vote),(total_votes_rep/total_vote)])
           
    fig.update_layout( title = {
        "text":"total votes for dem and rep",
        "y":0.95,
        "x":0.5
    },
            xaxis_title = "parties",
            yaxis_title = "% votes"
                 )
       
                 
    fig.show()
    most_voted_candidate()

画地图分析

对两个党派之间的数据进行整合 把胜利者的名字拎出来

president_election = pd.concat([rep_election,dem_election],axis=1)
president_election['winner'] = np.where(president_election.rep_votes>president_election.dem_votes,"Donald Trump","joe biden")
president_election                             

美国的各个州的经纬度表

path_states_map = "./data/statelatlong.csv"
states_lat_long = pd.read_csv(path_states_map,index_col="City")
states_lat_long

把经纬度表和美国选举的数据进行连接

president_election = pd.concat([president_election,states_lat_long],axis=1)
president_election

可以画出我们需要的地图了

fig = px.choropleth(
    president_election,
    locations="State",
    color="winner",
    color_discrete_sequence=["red","blue"],
    locationmode= "USA-states",
    scope = "north america",
    title ="USA Presidential Votes coun"
)
fig.show()

ok 做到这里就结束了这样做一个简单的项目了
需要这样子的小项目来练练手的话可以关注公众号来获取
输入【python_uaselect_001】获取数据集和相应的代码