Analysis of Fuel Price Data: Collection, Interpretation, and Statistical Inferences

University
Statistics and Data Analysis
By
Your Name
Date
Page 1 of 15
© <Your Name>2019

Section 1
Part A
Introduction
Information on Fuel price is usually collected and processed worldwide to assist the
government and the ministries involved in understanding the current trends in
prices of fuel from different brands, forecast the future patterns as well as
harmonize the prices amongst all the brands (Australian Competition and Consumer
Commission, 2019). It also helps in estimation of the ever growing economy of the
given country over a given period of time. Different methods of collection of the
data are applied to make sure that the data collected is adequate, reliable and
dependable over a given range of time (Hinton, 2014). These methods range from
direct surveys, using secondary sources as well as application of sophisticated
methods such as computer-based data collection and various analysis of data using
mathematical methods. However, while in the process of data collection and
interpretation, chances of biases affecting the data are usually introduced by the
method used in data collection or poor handling of the data. Such biases result to
inaccuracy and unreliability of the data (John and Nickel, 2013).
The objective of this paper is therefore, to examine and test the skills of collection
of both primary and secondary data, interpret the data and provide appropriate
numeric summaries, display the correct graphs and apply the necessary statistical
inferences that will aid the NRMA in making and presenting reliable and dependable
media report on fuel prices across various stations in different suburbs of Australia.
Page 2 of 15
© <Your Name>2019

Part B and C
Description of datasets
Dataset 1 can be described as a secondary data type since it is provided online by
the Australian Government Open Data under the license of Creative Commons
Attribution 3.0 Australia (Ullah, 2014). It is a subset of the Service Station and Price
History for the month of September 2016. This second hand or secondary data has
been edited to include only a few variables and cases to aid in answering specific
research question to provide necessary information for NRMA so they can deliver a
reliable media report on the prices of petrol. The data has seven variables. Service
Station name, Address, suburb, post code, brand and fuel code are all categorical
variables whereas the variable price is numerical (Bruce, 2015). The variable
service station name gives the name of the station, address gives the exact address
of the station, suburb shows the suburb in which the station is located, postcode
indicates the Australian postal code of the suburb, Brand indicates the brand of
service for a given station, fuel code indicates the type of fuel code while the price
variable indicates the price in Australian Cents for the petrol. The variable suburb
has cases such as Manly, Emu Plains etc., the variable brand has cases such as
Caltex, Caltex Woolworths, United, BP and Puma energy (Freund, 2014).
Dataset 2 is Primary data (Bissett, 2007). It is collected by direct survey of 30 KOI
colleagues on their most preferred service station from which they buy fuel. The
dataset has got three variables namely; date, gender, and address of the preferred
station. The variable date indicates the day on which the data was collected from
the interviewee, gender is a categorical variable that indicates the whether the
Page 3 of 15
© <Your Name>2019

respondent is either male or female while address is a categorical variable that
indicates the address of the station where the respondent prefers to buy the fuel
(Newbold, Carlson, and Thorne, 2013). The limitation of this type of data is that, the
sample comprising of the 30 KOI students; both male and female is somehow small
and therefore is not representative of the whole school’s population on their
preferred station of fuel (Shao, 2010). Additionally, considering that the students
have shared the same class for a long time, a friendship bond between them could
have created a possibility of the responses given being influenced by the
interviewees sharing the same interests or not being independent and free while
giving their responses. The accuracy and reliability of this dataset could be increase
by sampling more people from different classes and having different interests
(Tuffery, 2013).
Section 2
Analysis of Single Variable in dataset 1
a. Shape of the distribution of the Variable Price
To determine the shape of the distribution of the variable price for the
different stations we use the numerical summary table, frequency distribution
table and histogram (Croucher, 2016). The descriptive statistics include the
mean, mode, median, quartiles, range, interquartile range among others. The
numerical summary table is as shown below:
Page 4 of 15
© <Your Name>2019