import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns %matplotlib inline plt.rcParams['font.sans-serif'] = ['SimHei'] plt.rcParams['axes.unicode_minus'] = False plt.style.use('ggplot') df = pd.read_csv('train.csv') df.head(2) print(df.shape,'\n') df.info() df.describe() df['date'] = df['datetime'].apply(lambda x:x.split(' ')[0]) # 获取小时数 df["hour"] = df['datetime'].apply(lambda x: x.split()[1].split(":") [0]) del df['datetime'] # 获取星期数 df['day'] = df['date'].apply(lambda x:pd.to_datetime(x,format='%Y -%m-%d')).dt.weekday # 获取月份数 df['month'] = df['date'].apply(lambda x:pd.to_datetime(x,format='% Y-%m-%d')).dt.month # 将date 字段改为时间格式 df['date'] = pd.to_datetime(df['date'],format = '%Y-%m-%d') print(df.info()) df.head(2) # 工作日用车情况 plt.figure(figsize=(15,6)) df[df['workingday']==1].groupby('hour').mean()['count'].plot(line style = '-', # 非工作日用车情况 df[df['workingday']==0].groupby('hour').mean()['count'].plot(line style = '-', linewidth = 2, color = 'g') linewidth = 2, color = 'y') 

plt.xticks(range(24),df['hour'],fontsize=15) plt.yticks(fontsize=15) plt.legend(['工作日','非工作日'],fontsize=15) plt.title(u'一天各时段用车人数 (工作日 VS 非工作日)',fontsize=20) plt.show() plt.figure(figsize=(10,6)) plt.bar(range(7),df.groupby('day').mean()['count'],width=0.3) plt.ylim([180,200]) plt.title('一周内单车需求量',fontsize=20) plt.legend(['人数'],fontsize=15) for x,y in enumerate(df.groupby('day').mean()['count']): plt.text(x-0.12,y+0.3,'%i'%y,va='center',fontsize=15) plt.figure(figsize=(15,6)) df.groupby('month').mean()['count'].plot(marker = 'o') plt.xticks(range(0,13)) plt.title('不同月份用车情况对比',fontsize=20) plt.legend(['人数'],fontsize=15) plt.figure(figsize=(15,6)) # 2011 年 df[df['date'].map(lambda x:x.year)==2011].groupby('month')['count '].mean().plot() # 2012 年 df[df['date'].map(lambda x:x.year)==2012].groupby('month')['count '].mean().plot() plt.xticks(range(0,13)) plt.legend(['2011 年','2012 年'],fontsize=15) df_registered = df['registered'].sum() / df['count'].sum() df_casual = df['casual'].sum() / df['count'].sum() plt.figure(figsize=(15,6)) plt.axis('equal') plt.pie(x = [df_registered,df_casual], autopct = '%.1f%%', textprops = {'fontsize' : 15}, explode = [0.08,0], shadow=True) plt.legend(['会员','非会员'],fontsize=12,loc = 'upper left') plt.figure(figsize=(15,6)) df.groupby('hour')['registered'].mean().plot() df.groupby('hour')['casual'].mean().plot() 

plt.xticks(range(24),df['hour'],fontsize=15) plt.legend(['会员','非会员'],fontsize = 15) plt.title('24 小时单车使用情况 (会员 VS 非会员)',fontsize=20) plt.show() df_week_registered = df.groupby('day').mean()['registered'] df_week_casuual = df.groupby('day').mean()['casual'] plt.figure(figsize=(15,6)) df_week_registered.plot() df_week_casuual.plot() plt.legend(['会员','非会员']) plt.figure(figsize=(15,6)) df[df['season']==1].groupby('hour').mean()['count'].plot() df[df['season']==2].groupby('hour').mean()['count'].plot() df[df['season']==3].groupby('hour').mean()['count'].plot() df[df['season']==4].groupby('hour').mean()['count'].plot() plt.legend(['春','夏','秋','冬'],fontsize=14) plt.xticks(range(24),df['hour'],fontsize=15) plt.title('季节与用车量的关系',fontsize=20) plt.show()