QUA4A5 Assessment of Transport for London

QUA4A5 Assessment 3: Group Project
TfL Case study assignment
Name of the Student
Name of the University

Table of Contents
Introduction................................................................................................................................3
Task I..........................................................................................................................................3
Part [a]....................................................................................................................................3
Part [b]....................................................................................................................................3
Part [c]....................................................................................................................................4
Task II........................................................................................................................................5
Part [a]....................................................................................................................................5
Part [b]....................................................................................................................................5
Task III.......................................................................................................................................5
Part [a]....................................................................................................................................5
Part [b]....................................................................................................................................6
Task IV.......................................................................................................................................7
Part [a]....................................................................................................................................7
Part [b]....................................................................................................................................8
Task V........................................................................................................................................9
Part [a]....................................................................................................................................9
Part [b]....................................................................................................................................9
Part [c]..................................................................................................................................10
Conclusion................................................................................................................................10
Reference..................................................................................................................................11

Introduction
The aim of this project is to evaluate a large data gathered by Transport for London's (TfL's)
related to their 'Santander' Bike Hire Scheme. In specific, in this project work, being the
director of TfL, the author tried to investigate number of bike hires on each day, each season,
the general trend and any association between daily temperature and number of bike hires.
The entire investigation has been divided in 5 sub tasks that are mentioned in the subsequent
section. The author at the same time used MS excel function to predict 7 days possible
number of bikes hiring.
Task I
Part [a]
In this section, the author has calculated mean, median and standard deviation of daily bike
hire based on historical data. The table below is showing the details:
Mean 31705.8
Median 33112
Sample Stdev 11300.3
Standard Error of Mean 292.583
Z95% 1.96
Lower confidence Limit 31132.3
Upper confidence Limit 32279.2
Table 1: mean, median, std dev, confidence interval
[Source: calculation done in excel]
The above table is also showing the 95% confidence interval of population mean of daily
bike hiring. From the above table, it can be concluded that on an average daily basis 31706
bikes were hired from different places in London. At the same time, this table is also
indicating that 50% cases the number of bikes hired on daily basis was more than 33112.
Based on this sample data, the author further concluded that there is a 95% chance that in any
days the number of bikes will be hired will remain in between 31132 to 32279.
Part [b]
The sample data set related to number of bicycles hiring is not normally distributed. Rather,
with reference to the histogram mentioned below, it can be concluded that the sample is
following a bimodal distribution.

3593
8179.605263
12766.21053
17352.81579
21939.42105
26526.02632
31112.63158
35699.23684
40285.84211
44872.44737
49459.05263
54045.65789
58632.26316
0
20
40
60
80
100
120
Histogram
Frequency
Bin
Frequency
Figure 1: Histogram of daily bike hiring
[Source: MS excel output]
Since, the distribution is bimodal in nature, it can be said that there are broadly two different
groups, one of which is hiring significantly large number of bikes have been hired, when the
other group is indicating significantly less number of bikes have been hired.
Part [c]
A dataset is normally distributed or not can be expressed with two parameters such as mean
and standard deviation (Siegel, 2016). The parameter, mean is indicating how data from a
process is distributed, when the parameter, standard deviation indicates the spread of the data.
A normally distributed dataset is very much essential for data analysis and therefore
concluding the trend basis the dataset. Taken for example, if a dataset is not normally
distributed means it has outliers. In this context, if the mean of the dataset is taken for
consideration, then the existence of outlier will make the mean value either significantly high
or significantly low (Anderson et al. 2020). However, if such outliers are removed, then the
mean value will give an accurate result. Similarly, calculating moments, correlations between
variables, and other calculations that are domain specific. In such case, if the dataset is not
normally distributed, then there will be significant deviation from actual context to predicted
context.