读取数据集😄

import pandas as pd
import warnings

warnings.filterwarnings('ignore')

data = pd.read_csv('./dataset/metro-bike-share-trip-data.csv')

data.head(1)

	Trip ID	Duration	Start Time	End Time	Starting Station ID	Starting Station Latitude	Starting Station Longitude	Ending Station ID	Ending Station Latitude	Ending Station Longitude	Bike ID	Plan Duration	Trip Route Category	Passholder Type	Starting Lat-Long	Ending Lat-Long
0	1912818	180	2016-07-07T04:17:00	2016-07-07T04:20:00	3014.0	34.05661	-118.23721	3014.0	34.05661	-118.23721	6281.0	30.0	Round Trip	Monthly Pass	{'longitude': '-118.23721', 'latitude': '34.05...	{'longitude': '-118.23721', 'latitude': '34.05...

分析思路(Analysis methods)😜

热门Station分布分析(Distribution of Hot Stations)
2.共享单车出行高峰时间段分析(Rush Hour Analysis)
3.单程与往返的比较分析(Round Trip/One Way Comparison Analysis)
使用时长与会员卡类型的关联度分析(relevancy analysis between Duration&Passholder Type)

开始数据分析(Data Analysis)😁

1）共享单车热门开始站分布可视化探索

pic = st_info.plot(kind="scatter", x="Starting Station Longitude", y="Starting Station Latitude",s = st_info['Counts']/10,
                        alpha=1)
plt.axis([-118.28,-118.22,34.02,34.07])
plt.show()

Conclusion

If there is a L.A. map should be better to analysis...

2）共享单车出行高峰时间段分析(Rush Hour Analysis)

Conclusion

共享单车出行者通常选择在**7a.m.-20p.m.**出行
**6a.m.-8a.m.**早高峰时段，10a.m.-12a.m.午高峰时段以及15a.m.-17a.m.晚高峰时段都有明显的用车量增加。
出行最高峰是17p.m.，推测为下班晚高峰
出行最低谷为4a.m.，凌晨的用车辆非常少
从高峰到低谷是一个完全下降趋势，用车辆逐渐降低
Shared-bike riders usually choose to start their trip at 7a.m.-20p.m
**6a.m.-8a.m.**morning peak period, **10a.m.-12a.m.**afternoon peak period and **15a.m.-17a.m.**evening peak period shows significant increase in shared-bike use
The peak of use is 17p.m., which is presumed to be the evening peak
The lowest point of use is 4a.m,few shared-bikes are used in mid-night
From the peak to the lowest point is a complete downward trend, with the vehicle gradually reduced

3）单车使用时长与会员卡持有种类的相关性分析

(relevancy analysis between Duration&Passholder Type)

bike_trip_info = bike_info[['Duration','Trip Route Category','Plan Duration','Passholder Type']]
bike_trip_info.head()

	Duration	Trip Route Category	Plan Duration	Passholder Type
0	180	Round Trip	30.0	Monthly Pass
1	1980	Round Trip	30.0	Monthly Pass
2	300	Round Trip	365.0	Flex Pass
3	10860	Round Trip	365.0	Flex Pass
4	420	Round Trip	0.0	Walk-up

单独提取Duration，进行聚类(cluster by Duration)

duration = bike_trip_info[['Duration']]
duration.head()
# duration时间单位为 秒

	Duration
0	180
1	1980
2	300
3	10860
4	420

bike_trip_info.head()

	Duration	Duration Class	Trip Route Category	Plan Duration	Passholder Type
0	180	short-time	Round Trip	30.0	Monthly Pass
1	1980	short-time	Round Trip	30.0	Monthly Pass
2	300	short-time	Round Trip	365.0	Flex Pass
3	10860	medium-time	Round Trip	365.0	Flex Pass
4	420	short-time	Round Trip	0.0	Walk-up

查看各类Duration数量

bike_trip_info['Duration Class'].value_counts()

short-time        127398
medium-time         3180
long-time            582
very-long-time       501
Name: Duration Class, dtype: int64

Duration Class探索

雷达图可视化(Radar Chart)

Conclusion

共享单车使用使用时间特别长的用户中绝大部分使用者是即用即走型用户
短途共享单车使用者中超过6成比例为月卡持有者
其次，短途共享单车使用者中的年卡持有者的比例也是所有类型中最高的
The majority of users who use shared-bikes for a very-long-time are those who use them immediately.
Over 60% of users who use shared-bikes for a short-time are those who possess Monthly-Pass.
The percentage of those who possess Flex-Pass among short-time trip bike users is also the highest among all types.

4）单程与往返的比较分析(Round Trip/One Way Comparison Analysis)

划分出单程与往返的frame(Divide the 'Round Trip'&'One Way')

one_way_trip.head()

	Duration	Start Time	End Time	Start Hour	Trip Route Category	Plan Duration	Passholder Type
5	780	2016-07-07T12:51:00	2016-07-07T13:04:00	12	One Way	30.0	Monthly Pass
6	600	2016-07-07T12:54:00	2016-07-07T13:04:00	12	One Way	30.0	Monthly Pass
7	600	2016-07-07T12:59:00	2016-07-07T13:09:00	12	One Way	365.0	Flex Pass
9	960	2016-07-07T13:01:00	2016-07-07T13:17:00	13	One Way	30.0	Monthly Pass
10	960	2016-07-07T13:02:00	2016-07-07T13:18:00	13	One Way	365.0	Flex Pass

round_trip.head()

	Duration	Start Time	End Time	Start Hour	Trip Route Category	Plan Duration	Passholder Type
0	180	2016-07-07T04:17:00	2016-07-07T04:20:00	04	Round Trip	30.0	Monthly Pass
1	1980	2016-07-07T06:00:00	2016-07-07T06:33:00	06	Round Trip	30.0	Monthly Pass
2	300	2016-07-07T10:32:00	2016-07-07T10:37:00	10	Round Trip	365.0	Flex Pass
3	10860	2016-07-07T10:37:00	2016-07-07T13:38:00	10	Round Trip	365.0	Flex Pass
4	420	2016-07-07T12:51:00	2016-07-07T12:58:00	12	Round Trip	0.0	Walk-up

分析One Way与Round出发时间点

分析One Way与Round中不同会员卡持有者比例

Conclusion

One Way Trip的共享单车使用者持有会员卡的比例很大，占了超过70%，其中绝大部分是月卡持有者。
Round Trip的共享单车使用者大部分是即来即走型使用者，仅有三成用户持有会员卡，持有会员卡的用户中大部分为月卡持有者。

分析One Way与Round与共享单车使用时长的关系

duration_cate_info = bike_info[['Duration','Trip Route Category']]
duration_cate_info.head()

	Duration	Trip Route Category
0	180	Round Trip
1	1980	Round Trip
2	300	Round Trip
3	10860	Round Trip
4	420	Round Trip

输出describe信息

	One Way	Round
count	119026.000000	12635.000000
mean	1358.870499	3299.287693
std	5490.783118	7738.171315
min	60.000000	60.000000
25%	360.000000	900.000000
50%	600.000000	1680.000000
75%	960.000000	3180.000000
max	86400.000000	86400.000000

Conclusion

根据上表中均值，中位数，上下四分位点可以显然得知Round Trip单车使用时长往往大于One Way Trip的使用时长，符合认知。

About

基于Kaggle数据集“洛杉矶共享单车数据”进行的数据探索与分析。

Languages

Language:Python 100.0%