如何用python做出超炫酷图表？

CDA数据分析研究院原创作品，转载需授权小编总是被那些玩转数据、利用数据做出超炫酷图表的大佬深深折服，膝盖都不够给他们。进行数据可视化做出超炫图表的软件有很多，今天小编也用数据分析常用的python来演示一下如何做出精彩的数据可视化呈现。导入相关的库和加载数据import numpy as np import pandas as pd import seaborn as snsimport matplotlib.pyplot as pltfrom datetime import date, timedelta, datetime设置路径和加载数据小编使用的是一个记录美国1908年到2009年飞机出事和死亡乘客记录的数据。import osos.chdir(r’D:dataair_data’)Data=pd.read_csv(‘airplane.csv’)查看各列有没有缺失值：Data.isnull().sum()对缺失数据进行清洗：Data[‘Time’] = Data[‘Time’].replace(np.nan, ‘00:00’) Data[‘Time’] = Data[‘Time’].str.replace(‘c: ‘, ‘’)Data[‘Time’] = Data[‘Time’].str.replace(‘c:’, ‘’)Data[‘Time’] = Data[‘Time’].str.replace(‘c’, ‘’)Data[‘Time’] = Data[‘Time’].str.replace(‘12'20’, ‘12:20’)Data[‘Time’] = Data[‘Time’].str.replace(‘18.40’, ‘18:40’)Data[‘Time’] = Data[‘Time’].str.replace(‘0943’, ‘09:43’)Data[‘Time’] = Data[‘Time’].str.replace(‘22'08’, ‘22:08’)Data[‘Time’] = Data[‘Time’].str.replace(‘114:20’, ‘00:00’)Data[‘Time’] = Data[‘Date’] + ’ ’ + Data[‘Time’]return datetime.strptime(x, ‘%m/%d/%Y %H:%M’)Data[‘Time’] = Data[‘Time’].apply(todate) print(‘Date ranges from ’ + str(Data.Time.min()) + ’ to ’ + str(Data.Time.max()))Data.Operator = Data.Operator.str.upper() 数据可视化绘制1908年到2009年飞机出事频数的折线图，大概得出一个趋势变化。Temp = Data.groupby(Data.Time.dt.year)[[‘Date’]].count()Temp = Temp.rename(columns={“Date”: “Count”})plt.figure(figsize=(12,6))plt.style.use(‘bmh’)plt.plot(Temp.index, ‘Count’, data=Temp, color=‘blue’, marker = “.”, linewidth=1)plt.xlabel(‘Year’, fontsize=10)plt.ylabel(‘Count’, fontsize=10)plt.title(‘Count of accidents by Year’, loc=‘Center’, fontsize=14)plt.show()我们把时间再精细化点，观察每月，每个星期，甚至每小时的事故，这次我们不看趋势，看量，绘制条形图。import matplotlib.pylab as plimport matplotlib.gridspec as gridspecgs = gridspec.GridSpec(2, 2)pl.figure(figsize=(15,10))plt.style.use(‘seaborn-muted’)ax = pl.subplot(gs[0, :]) # row 0, col 0sns.barplot(Data.groupby(Data.Time.dt.month)[[‘Date’]].count().index, ‘Date’, data=Data.groupby(Data.Time.dt.month)[[‘Date’]].count(), color=‘lightskyblue’, linewidth=2)plt.xticks(Data.groupby(Data.Time.dt.month)[[‘Date’]].count().index, [‘Jan’, ‘Feb’, ‘Mar’, ‘Apr’, ‘May’, ‘Jun’, ‘Jul’, ‘Aug’, ‘Sep’, ‘Oct’, ‘Nov’, ‘Dec’])plt.xlabel(‘Month’, fontsize=10)plt.ylabel(‘Count’, fontsize=10)plt.title(‘Count of accidents by Month’, loc=‘Center’, fontsize=14)ax = pl.subplot(gs[1, 0])sns.barplot(Data.groupby(Data.Time.dt.weekday)[[‘Date’]].count().index, ‘Date’, data=Data.groupby(Data.Time.dt.weekday)[[‘Date’]].count(), color=‘lightskyblue’, linewidth=2)plt.xticks(Data.groupby(Data.Time.dt.weekday)[[‘Date’]].count().index, [‘Mon’, ‘Tue’, ‘Wed’, ‘Thu’, ‘Fri’, ‘Sat’, ‘Sun’])plt.xlabel(‘Day of Week’, fontsize=10)plt.ylabel(‘Count’, fontsize=10)plt.title(‘Count of accidents by Day of Week’, loc=‘Center’, fontsize=14)ax = pl.subplot(gs[1, 1])sns.barplot(Data[Data.Time.dt.hour != 0].groupby(Data.Time.dt.hour)[[‘Date’]].count().index, ‘Date’, data=Data[Data.Time.dt.hour != 0].groupby(Data.Time.dt.hour)[[‘Date’]].count(),color =‘lightskyblue’, linewidth=2)plt.xlabel(‘Hour’, fontsize=10)plt.ylabel(‘Count’, fontsize=10)plt.title(‘Count of accidents by Hour’, loc=‘Center’, fontsize=14)plt.tight_layout()plt.show()出事时，每年登机人数与死亡人数的对比图Fatalities = Data.groupby(Data.Time.dt.year).sum()Fatalities[‘Proportion’] = Fatalities[‘Fatalities’] / Fatalities[‘Aboard’]plt.figure(figsize=(15,6))plt.subplot(1, 2, 1)plt.fill_between(Fatalities.index, ‘Aboard’, data=Fatalities, color=“skyblue”, alpha=0.2)plt.plot(Fatalities.index, ‘Aboard’, data=Fatalities, marker = “.”, color=“Slateblue”, alpha=0.6, linewidth=1)plt.fill_between(Fatalities.index, ‘Fatalities’, data=Fatalities, color=“olive”, alpha=0.2)plt.plot(Fatalities.index, ‘Fatalities’, data=Fatalities, color=“olive”, marker = “.”, alpha=0.6, linewidth=1)plt.legend(fontsize=10)plt.xlabel(‘Year’, fontsize=10)plt.ylabel(‘Amount of people’, fontsize=10)plt.title(‘Total number of people involved by Year’, loc=‘Center’, fontsize=14)plt.subplot(1, 2, 2)plt.plot(Fatalities.index, ‘Proportion’, data=Fatalities, marker = “.”, color = ‘red’, linewidth=1)plt.xlabel(‘Year’, fontsize=10)plt.ylabel(‘Ratio’, fontsize=10)plt.title(‘Fatalities / Total Ratio by Year’, loc=‘Center’, fontsize=14)plt.tight_layout()plt.show()通过对比图我们可以看到死亡人数变得如此之高(即使在90年代后似乎有下降的趋势)。一些人提出了一个很好的观点，那就是图表并没有显示每年所有航班发生事故的比例。因此，1970-1990年在空中交通信号灯的历史上看起来是可怕的一年，死亡人数上升，但也有可能是乘飞机的总人数上升，而实际上比例下降了。亲爱的筒子们，想了解更多用python玩转数据、掌握炫酷可视化技能那就赶紧关注CDA数据分析师wechat公众号(cdacdacda)吧，点赞、转发、收藏，更多干货内容呈现给你噢。https://www.cda.cn/kecheng/in…