A城市巡游车与网约车运营特征对比分析—地图及订单数据统计初始化地图数据统计：订单数据统计

初始化

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt 
import folium 

# 文件目录，绝对路径
INPUT_PATH = 'G:DCICData\'

# 文件读取行数
MAX_ROWS = 100000 

地图数据统计：

taxigps2019 = pd.read_csv(INPUT_PATH + 'taxiGps20190531.csv',
                         dtype = {
                             'DRIVING_DIRECTION': np.uint16,
                             'OPERATING_STATUS': np.uint8,
                             'LONGITUDE': np.float32,
                             'LATITUDE': np.float32,
                             'GPS_SPEED': np.float32 
                         })

taxigps2019 = taxigps2019[taxigps2019.columns[::-1]]
taxigps2019['GPS_TIME'] = pd.to_datetime(taxigps2019['GPS_TIME'])
taxigps2019.sort_values(by=['CARNO','GPS_TIME'], inplace=True)
taxigps2019.reset_index(inplace=True, drop=True)
taxigps2019.head() 

统计每辆巡游车最早、最晚出现的记录：

df_first = taxigps2019.groupby(['CARNO']).first()
print('最早出现的记录：')
df_first.head() 

最早出现的记录：

df_last = taxigps2019.groupby(['CARNO']).last()
print('最晚出现的记录：')
df_last.head() 

最晚出现的记录：

统计每辆巡游车最早最晚的时间间隔：

df_first = taxigps2019.groupby(['CARNO']).first()
df_last = taxigps2019.groupby(['CARNO']).last()

df = df_last['GPS_TIME'] - df_first['GPS_TIME']
df = df.reset_index()

df['GPS_HOUR'] = df['GPS_TIME'].dt.seconds / 3600 # 秒转换为小时
df['GPS_HOUR'] = df['GPS_HOUR'].astype(int)
df.set_index('CARNO', inplace=True)
df.head() 

统计每辆巡游车的经纬度和速度极差：

taxigps2019 = taxigps2019[taxigps2019['LATITUDE'] != 0]
taxigps2019 = taxigps2019[taxigps2019['LONGITUDE'] != 0]

df['LATITUDE_PTP'] = taxigps2019.groupby(['CARNO'])['LATITUDE'].apply(np.ptp)
df['LONGITUDE_PTP'] = taxigps2019.groupby(['CARNO'])['LONGITUDE'].apply(np.ptp)
df['GPS_SPEED_PTP'] = taxigps2019.groupby(['CARNO'])['GPS_SPEED'].apply(np.ptp)

df.head()

df[(df['LATITUDE_PTP'] == 0)&(df['LONGITUDE_PTP'] == 0)&(df['GPS_SPEED_PTP'] == 0)].count()

GPS_TIME         39
GPS_HOUR         39
LATITUDE_PTP     39
LONGITUDE_PTP    39
GPS_SPEED_PTP    39
dtype: int64

通过统计经纬度以及GPS速度的极差是否为0，可以剔除全天GPS异常的车辆。

计算每辆巡游车的平均经纬度，并绘制热力图：

df['LONGITUDE_MEAN'] = taxigps2019.groupby(['CARNO'])['LONGITUDE'].mean()
df['LATITUDE_MEAN'] = taxigps2019.groupby(['CARNO'])['LATITUDE'].mean()
df = df.dropna()

from folium import plugins
from folium.plugins import HeatMap

map_hooray = folium.Map(location=[24.482426, 118.157606], zoom_start=14)
HeatMap(df[['LATITUDE_MEAN', 'LONGITUDE_MEAN']].iloc[:1000].values).add_to(map_hooray)
map_hooray  

对比分析2019，2020年端午节前一天巡游车平均速度的变化：

# 2019年端午节6月7日，2020年端午节6月25日
taxigps2019 = pd.read_csv(INPUT_PATH + 'taxiGps20190606.csv',
                         dtype = {
                             'DRIVING_DIRECTION': np.uint16,
                             'OPERATING_STATUS': np.uint8,
                             'LONGITUDE': np.float16,
                             'LATITUDE': np.float16,
                             'GPS_SPEED': np.float32 
                         })
dwq2019 = taxigps2019['GPS_SPEED'].mean()
del taxigps2019
taxigps2020 = pd.read_csv(INPUT_PATH + 'taxiGps20200624.csv',
                         dtype = {
                             'DRIVING_DIRECTION': np.uint16,
                             'OPERATING_STATUS': np.uint8,
                             'LONGITUDE': np.float16,
                             'LATITUDE': np.float16,
                             'GPS_SPEED': np.float32 
                         })
dwq2020 = taxigps2020['GPS_SPEED'].mean()
del taxigps2020
print(dw2019)
print(dw2020) 

243.19984
218.5537

对比分析2019年端午假期前、端午假期中和假期后，巡游车日平均速度变化：

l_name = ['0606','0607','0608','0609']
speed2019 = []
for i in l_name:
    taxigps2019 = pd.read_csv(INPUT_PATH + 'taxiGps2019'+i+'.csv',
                         dtype = {
                             'DRIVING_DIRECTION': np.uint16,
                             'OPERATING_STATUS': np.uint8,
                             'LONGITUDE': np.float16,
                             'LATITUDE': np.float16,
                             'GPS_SPEED': np.float32 
                         })
    speed2019.append(taxigps2019['GPS_SPEED'].mean())

plt.plot(l_name,speed2019) 

订单数据统计

del taxigps2019
taxiorder2019 = pd.read_csv(INPUT_PATH + 'taxiOrder20190531.csv',
                           dtype = {
                               'GETON_LONGITUDE': np.float32,
                               'GETON_LATITUDE': np.float32,
                               'GETOFF_LONGITUDE': np.float32,
                               'GETOFF_LATITUDE': np.float32,
                               'PASS_MILE': np.float16,
                               'NOPASS_MILE': np.float16,
                               'WAITING_TIME': np.float32
                           })
taxiorder2019 = taxiorder2019.rename(columns={'CAR_NO':'CARNO'})
taxiorder2019.sort_values(by=['CARNO','GETON_DATE'], inplace=True) # 根据车号和上车时间进行排序
taxiorder2019.reset_index(inplace=True, drop=True)
taxiorder2019.head()

上下客点分布密度：

• 上车经纬度分布热力图：

df = taxiorder2019.groupby(['CARNO'])['GETON_LONGITUDE'].mean() 
df = df.reset_index()
df.set_index('CARNO', inplace=True)
df['GETON_LATITUDE'] = taxiorder2019.groupby(['CARNO'])['GETON_LATITUDE'].mean()
df.head() 

map_hooray = folium.Map(location=[24.482426, 118.157606], zoom_start=14)
HeatMap(df[['GETON_LATITUDE', 'GETON_LONGITUDE']].values).add_to(map_hooray)
map_hooray  

• 下车经纬度分布热力图：

df['GETOFF_LONGITUDE'] = taxiorder2019.groupby(['CARNO'])['GETOFF_LONGITUDE'].mean()
df['GETOFF_LATITUDE'] = taxiorder2019.groupby(['CARNO'])['GETOFF_LATITUDE'].mean()
df.head() 

map_hooray = folium.Map(location=[24.482426, 118.157606], zoom_start=14)
HeatMap(df[['GETOFF_LATITUDE', 'GETOFF_LONGITUDE']].values).add_to(map_hooray)
map_hooray  

巡游车在（118.155060±0.01）、（24.506035±0.01）方位内打车的平均等待时间：

taxiorder2019[(118.155060-0.01<taxiorder2019['GETON_LONGITUDE'])&(taxiorder2019['GETON_LONGITUDE']<118.155060+0.01)&
              (24.506035-0.01<taxiorder2019['GETON_LATITUDE'])&(taxiorder2019['GETON_LATITUDE']<24.506035+0.01)]['WAITING_TIME'].mean()

276.4932

统计出租订单平均等待时间超过600秒的经纬度：

f = lambda x:str(round(x,3))
taxiorder2019['GETON_POSITION'] = taxiorder2019['GETON_LONGITUDE'].apply(f)+','+taxiorder2019['GETON_LATITUDE'].apply(f)
df = taxiorder2019.groupby(['GETON_POSITION'])['WAITING_TIME'].mean()
df = df.reset_index()
df.set_index('GETON_POSITION', inplace=True) 
df = df[ taxiorder2019.groupby(['GETON_POSITION'])['WAITING_TIME'].mean() > 600]
df = df.reset_index() 
df['GETON_POSITION'].value_counts() 

118.147,24.456    1
118.158,24.492    1
117.997,24.615    1
118.158,24.488    1
118.122,24.621    1
                 ..
118.091,24.523    1
118.149,24.489    1
118.186,24.505    1
118.074,24.612    1
118.156,24.452    1
Name: GETON_POSITION, Length: 631, dtype: int64

统计20190531 - 20190609 期间出租订单经纬度上平均等待时间长的位置（且位置出现频率大于5）：

l_name = ['0531','0601','0602','0603','0604','0605','0606','0607','0608','0609']
from collections import Counter
counter = Counter()
for i in l_name:
    taxiorder2019 = pd.read_csv(INPUT_PATH + 'taxiOrder2019'+i+'.csv',
                           dtype = {
                               'GETON_LONGITUDE': np.float32,
                               'GETON_LATITUDE': np.float32,
                               'GETOFF_LONGITUDE': np.float32,
                               'GETOFF_LATITUDE': np.float32,
                               'PASS_MILE': np.float16,
                               'NOPASS_MILE': np.float16,
                               'WAITING_TIME': np.float32
                           })
    taxiorder2019 = taxiorder2019.rename(columns={'CAR_NO':'CARNO'})
    taxiorder2019.sort_values(by=['CARNO','GETON_DATE'], inplace=True) # 根据车号和上车时间进行排序
    taxiorder2019.reset_index(inplace=True, drop=True)
    taxiorder2019['GETON_POSITION'] = taxiorder2019['GETON_LONGITUDE'].apply(f)+','+taxiorder2019['GETON_LATITUDE'].apply(f)
    df = taxiorder2019.groupby(['GETON_POSITION'])['WAITING_TIME'].mean()
    df = df.reset_index()
    df.set_index('GETON_POSITION', inplace=True) 
    df = df[ taxiorder2019.groupby(['GETON_POSITION'])['WAITING_TIME'].mean() > 600]
    df = df.reset_index() 
    counter.update(df['GETON_POSITION'].values) 

sorted(counter.items(),key=lambda x:x[1], reverse = True) 

[('0.0,0.0', 10),
 ('117.939,24.492', 8),
 ('118.13,24.486', 7),
 ('118.023,24.48', 6),
 ('118.109,24.431', 6),
 ('118.226,24.569', 6),
 ('118.019,24.482', 5),
 ('118.052,24.5', 5),
 ('118.166,24.481', 5),
 ('118.167,24.48', 5),
 ('118.228,24.568', 5),
 ('118.11,24.489', 5),
 ('118.036,24.491', 5),
 ('118.068,24.464', 5),
 ...]

对比2019年和2020年出租车端午节订单的平均等待时间，是如何变化的：

# 2019年端午节6月7日，2020年端午节6月25日
taxiorder2019 = pd.read_csv(INPUT_PATH + 'taxiOrder2019'+'0607'+'.csv',
                           dtype = {
                               'GETON_LONGITUDE': np.float32,
                               'GETON_LATITUDE': np.float32,
                               'GETOFF_LONGITUDE': np.float32,
                               'GETOFF_LATITUDE': np.float32,
                               'PASS_MILE': np.float16,
                               'NOPASS_MILE': np.float16,
                               'WAITING_TIME': np.float32
                           })
taxiorder2019 = taxiorder2019.rename(columns={'CAR_NO':'CARNO'})

taxiorder2020 = pd.read_csv(INPUT_PATH + 'taxiOrder2020'+'0625'+'.csv',
                           dtype = {
                               'GETON_LONGITUDE': np.float32,
                               'GETON_LATITUDE': np.float32,
                               'GETOFF_LONGITUDE': np.float32,
                               'GETOFF_LATITUDE': np.float32,
                               'PASS_MILE': np.float16,
                               'NOPASS_MILE': np.float16,
                               'WAITING_TIME': np.float32
                           })
taxiorder2020 = taxiorder2020.rename(columns={'CAR_NO':'CARNO'})

dw2019 = taxiorder2019['WAITING_TIME'].mean()
dw2020 = taxiorder2020['WAITING_TIME'].mean()
print(dw2019)
print(dw2020)

243.19984
218.5537

巡游车订单距离与时间的对比：

import matplotlib.animation as animation
import glob, gc

def barlist(n): 
    taxiorder2019 = pd.read_csv(paths[n], nrows=None,
                                   dtype = {
                                       'GETON_LONGITUDE': np.float32,
                                       'GETON_LATITUDE': np.float32,
                                       'GETOFF_LONGITUDE': np.float32,
                                       'GETOFF_LATITUDE': np.float32,
                                       'PASS_MILE': np.float16,
                                       'NOPASS_MILE': np.float16,
                                       'WAITING_TIME': np.float16
                                   })
    taxiorder2019['GETON_DATE'] = pd.to_datetime(taxiorder2019['GETON_DATE'])
    taxiorder2019['GETON_Hour'] = taxiorder2019['GETON_DATE'].dt.hour
    return taxiorder2019.groupby(['GETON_Hour'])['PASS_MILE'].mean().values # 不同上车时间的平均行驶距离

fig=plt.figure()

paths = glob.glob(INPUT_PATH +'taxiOrder20190*.csv')
paths.sort()
n = len(paths) #Number of frames
x = range(24)
barcollection = plt.bar(x,barlist(0))
plt.ylim(0,8)

def animate(i):
    y=barlist(i+1)
    for idx, b in enumerate(barcollection):
        b.set_height(y[idx])
    plt.ylim(0,8)
    plt.title(paths[i+1].split('/')[-1])
    plt.ylabel('PASS_MILE / KM')
    plt.xlabel('Hour')

anim=animation.FuncAnimation(fig,animate,repeat=False,blit=False,frames=n-1,
                             interval=500)

anim.save('order.gif', dpi=150)

巡游车空驶率对比：

for path in paths:
    taxiorder2019 = pd.read_csv(path, nrows=None,
                               dtype = {
                                   'GETON_LONGITUDE': np.float32,
                                   'GETON_LATITUDE': np.float32,
                                   'GETOFF_LONGITUDE': np.float32,
                                   'GETOFF_LATITUDE': np.float32,
                                   'PASS_MILE': np.float16,
                                   'NOPASS_MILE': np.float16,
                                   'WAITING_TIME': np.float16
                               })

    taxiorder2019 = taxiorder2019[['NOPASS_MILE', 'PASS_MILE']].dropna()
    taxiorder2019['NOPASS_Ratio'] = taxiorder2019['NOPASS_MILE'] / (taxiorder2019['NOPASS_MILE'] + taxiorder2019['PASS_MILE'])
    print(path, taxiorder2019['NOPASS_Ratio'].astype(np.float32).mean()) 

G:DCICDatataxiOrder20190531.csv 0.27126783
G:DCICDatataxiOrder20190601.csv 0.27297953
G:DCICDatataxiOrder20190602.csv 0.30302802
G:DCICDatataxiOrder20190603.csv 0.31049386
G:DCICDatataxiOrder20190604.csv 0.3039471
G:DCICDatataxiOrder20190605.csv 0.2933384
G:DCICDatataxiOrder20190606.csv 0.2547359
G:DCICDatataxiOrder20190607.csv 0.28453994
G:DCICDatataxiOrder20190608.csv 0.304996
G:DCICDatataxiOrder20190609.csv 0.3115026

最后

以上就是暴躁溪流为你收集整理的A城市巡游车与网约车运营特征对比分析—地图及订单数据统计初始化地图数据统计：订单数据统计的全部内容，希望文章能够帮你解决A城市巡游车与网约车运营特征对比分析—地图及订单数据统计初始化地图数据统计：订单数据统计所遇到的程序开发问题。

如果觉得靠谱客网站的内容还不错，欢迎将靠谱客网站推荐给程序员好友。