python笔记（二）一些有用的代码

热力图 heatmap

import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns
from pylab import *
import matplotlib.cm as cm
from matplotlib.colors import LogNorm

一列数画图

df = df.pivot("hour", "minute", "rate")
ax = sns.heatmap(df,annot= False,fmt='d', linewidths=.5, cmap='YlGnBu')
plt.show()

方阵（xy任意标注名称问题未解决）

data = np.array(df)

#sns.heatmap(data,annot= False,annot_kws={'size':9,'weight':'bold', 'color':'w'},fmt='.2f')#true 显示数值
sns.heatmap(data,annot= False,fmt='d', linewidths=.5, cmap='YlGnBu')
plt.show()

方阵

df = pd.read_csv(r'D:任选的一天1120方阵.csv',header = None)
df=df.astype(float)
#sns.heatmap(df)
sns.heatmap(df,cmap='Blues',center=-0.8)#颜色，中心对齐值，

plt.show()

数据清洗
excel可做的处理
分列（根据分隔符 / 同格式时 按长度分）
选择指定数据所在的行，形成新的dataframe

newdf1 = df[df[0] == 'A1.GT.24538' ]
print(newdf1)

去除重复项

newdf1 = newdf1.drop_duplicates()

缺失值查找（以日期为例）

for i in range (19,31):
    i=str(i)
    print(i)
    #date = date.replace(i,i+1)
    datei=''
    date=('2019-11-', i,'T')
    datei = datei.join(date)
    #print(datei)
    checki = newdf1.loc[newdf1[1] ==datei, :]
    print (checki.shape)

把日期分成年月日
并按礼拜分为工作日和非工作日

    # 日期
    date = df.values[:,1]
    hour_minutes = df.values[:,2]
    rate = df.values[:, 8]
    year = []
    month = []
    day = []
    hour = []
    minutes = []
    weekday = []
    up5 = []
    up10 = []
    up15 = []

    for i in range(3, len(date)):
        year.append(int(date[i][0:4]))
        month.append(int(date[i][5:7]))
        day.append(int(date[i][8:10]))
        # print(month)

    for i in range(0, len(date)-3):
        y = year[i]
        m = month[i]
        d = day[i]
        lis = ['1', '2', '3', '4', '5', '6', '7', ]
        dic = dict(enumerate(lis))
        #if y.isdigit() and m.isdigit() and d.isdigit() and 1 <= int(m) <= 12 and 1 <= int(d) <= 31:
        w = datetime.date(int(y), int(m), int(d))

        weekday.append(dic[w.weekday()])
    #print(len(weekday))
    print(len(weekday))

    weekday1 = []
    weekend1 = []
    weekday = list(map(int, weekday))
    #print(type(weekday[1]))


    for i in range(len(weekday)):
         if (weekday[i] == 6) or (weekday[i] == 7):
        #if  weekday[i]  == 6 :
             weekend1.append(df.values[i,:])
        else:
            weekday1.append(df.values[i,:])

转为dataframe格式
转置
导出

    weekday2 = pd.DataFrame([weekday1])
    weekday2 = weekday2.T
    weekday2.to_csv(r'D:\seu-monash\卢高速\weekday.csv')