Assignment

(Data Analytics and Business Intelligence /PG – 8696/8697)

This assessment item contributes 35% of the total marks of this unit and is due in Week 12
on Friday at 23:59. You should plan and complete your assignment before the deadline. The
deadline is the local time of Canberra, Australia. It is your responsibility to correctly adjust
your clocks, including the one in Canvas personalized user interface.

Submissions (single MS Word or pdf document) are through Canvas

(http://uclearn.canberra.edu.au/). Please familiarize yourself with Canvas submissions
before the deadline. After you upload your assignment report, please DO NOT FORGET to
click the submit button; otherwise, your submitted assignment may remain as a draft.

No other forms of submission will be accepted.

Extensions

Students can apply for an extension to the submission due date for an assessment item
through extenuating, evidenced circumstances (specific details are found through the
Assessment Policy and Procedures. Section 9.12). Extensions must be applied for before
the due date. Documentary evidence (e.g., medical certificate) will be expected for an
extension to be granted, however this will not guarantee that the application will be
successful. The Unit Convener or relevant Discipline Convener will decide whether to
grant an extension and the length of the extension.

An Assignment Extension form is available from the Student Forms page.

Late submission of assignments without an approved extension will result in a penalty

of 5% reduced marks from the total available, per calendar day late. An assignment
submitted over 7 days late will not be accepted.

If a student chooses to submit his/her assignment via the Internet off the campus, it is
the responsibility of the student to guarantee the accessibility of the Internet. Not being
able to access to the Internet at a location which is outside of the campus is not an
excuse for extension.

This assessment is a group task and only one submission per group is allowed. Ensure
all resources used are appropriately referenced in a referencing style of your choice;
but you need to maintain consistency.

Please submit your draft report to URKUND Student Text-matching Checker for a
similarity/plagiarism check before you submit your final report.

1
Overview
In this assignment you will be working with real-world data and problems. You will be
required to analyse a climate/weather data discussed on page 4. You are expected to
perform five of the six steps in the data mining process on your chosen dataset. These five
steps are:

• Business understanding
• Data understanding
• Data preparation
• Modelling
• Evaluation

Image: http://www.proglobalbusinesssolutions.com/six-steps-in-crisp-dm-the-standard-data-mining-process

As it is the case for most real-world data mining projects, each of the datasets requires
some pre-processing to get the data into a form suitable for mining.

2
Problem formulation
In this assignment, you are going to work on applying data mining to solve a real-world
business problem. This involves identifying a business in Australia that is impacted by
rainfall. The impact of rainfall on a business could be positive or negative.

• If the impact of rainfall on the business is positive, it creates an opportunity. The aim
is to make recommendations that help the business capitalise on the opportunity
created by rainfall.
• If the impact of rainfall on the business if negative it poses a risk, and the business
may incur loss. The aim is to make recommendations that help the business
minimise the loss incurred due to rainfall.

The first step will be to identify a business that is likely to be impacted (positively or
negatively) by rainfall and obtain data on their business operations such as revenue over a
period of time. Once the revenue (or information that will help us quantify the impact of
rainfall on the business) information is obtained, the weather data for the business location
(closest) can be downloaded from the Bureau of Meteorology website to explore if there is
correlation between revenue of the business and rainfall.

If a correlation (positive or negative) exists between business revenue and rainfall, data
mining may be used to build predictive models to forecast rainfall and make
recommendations to improve its business processes.

Alternative problems and tools

You are allowed and encouraged to formulate your own project problem in other areas of
applications of data mining. You can obtain and use datasets other than what is provided
in this assignment as long as you can formulate a data mining problem that requires you
to apply the CRISP-DM steps.

You are also allowed to perform all tasks using R or Python if you are proficient in using
those programming languages. You will need to submit your fully commented scripts
along with your final report.

Presentation of the progress report

As shown in the figure above, a data mining process involves six main steps. It is estimated
Data Preparation (understanding and cleaning data) takes 50% – 80% of the time spent on a
data mining project. It is strongly recommended that understanding of business and data is
completed and cleaning of data has commenced at this milestone.

Presentation of the progress report (PowerPoint presentation in Week 7) provides an

opportunity to discuss any challenges you are facing as a group in undertaking the project. It
also allows the unit convenor to assess the progress that has been made and provide
direction to help complete the project successfully and timely.

3
Deliverables
The deliverable for this assignment will be a report with associated datasets in csv file
format and fully commented scripts (R, Python), if any. The report will be in the style of a
professional document of no more than 12 pages1 single-spaced A4, Times New Roman
(12), including relevant graphs and output to support your reporting. A suggested
breakdown of the 12 pages is given in the sections below. However, you may apply a
different breakdown as needed. The report will detail your methods and the results of your
modelling. The report will contain the following sections:

1. Executive Summary (1 page/5 marks)

o Summarise the data, methods and results of your report.
2. Introduction (1 page/5 marks)
o Describe the context of the problem for a geographical area of your choice.
o Why is this problem important for a geographical area of your choice?
o What are the aims of the modelling for the business of your choice?
3. Data (2 pages/5 marks)
o Describe your data visually and numerically.
o Provide details of the pre-processing undertaken in order to meet the
conditions of your chosen modelling techniques.
4. Analyses (2 pages/5 marks)
o Describe the modelling procedure with enough detail to allow someone else
to repeat the modelling. At least two (for UG level students) and three (for
PG level students) different modelling techniques (e.g., decision trees,
ensemble decision trees) should be applied to the problem.
5. Results (2 pages/5 marks)
o Describe the performance of your models. Use appropriate charts and
visualisation to convey the results to your audience.
6. Discussion (2 pages/5 marks)
o Interpret your results.
o Which model performed the best? Why?
o What do the results mean in light of your aims?
o Which were the most important variables in your model?
7. Conclusion (1 page/5 marks)
o Summarise your findings.
8. Citation and references (1 page/no marks but required)
o Use Harvard or APA author-date style e.g., Richardson (2012, p. 217)
suggested that …

Note: Results and discussions may also be presented as a combined section.

1
Prior approval from the convenor is required to exceed the page limit. Exceeding page limit without
approval may result in a penalty of up to 5%.

4
Datasets

Weather dataset

This data can be retrieved online for a geographical area of your choice, from
http://www.bom.gov.au/climate/data/index.shtml?bookmark=200

Download each of the last 14 months of data (you may have to download these in 14
separate files).

Once the data has been downloaded, create a single spreadsheet by combining the 14
separate files you have downloaded. You will then need to use Excel to add a derived target
variable – RainTomorrow (Hint: its values may be determined by applying a threshold on the
Rainfall variable).

This historic dataset can be used to build a predictive model to predict whether it will rain
tomorrow. Decision tree is one modelling technique that may be appropriate for this
problem; you will need to choose and apply at least two different (three for PG students)
modelling techniques.

5
Marking rubric for the report

Marking Rubric

Pass (50% – 64%) Credit (65% – 74%) Distinction (75% – 84%) HD (85% – 100%)

Executive Basic Competent Proficient Advanced

summary Goal/hypothesis Goal/hypothesis Briefly describes Briefly describes
incomplete. mentioned but unclear or goal/hypothesis of goal/hypothesis of
confusing. modelling. modelling.

Little context provided, Context is provided, but Provides (in a few Provides (in a few sentences)
or context unclear or links to the sentences) the context of the context of problem being
confusing. goal/hypothesis unclear. problem being explored. explored. Clearly identifies
why the reader should care
about the problem.

Summary of results and Provides brief summary of Provides brief summary of

conclusion unclear. results and conclusion. methods employed, results
obtained, and conclusion
reached.

Introduction Basic Competent Proficient Advanced

Goal/hypothesis Goal/hypothesis Describes goal/hypothesis Describes goal/hypothesis of

6
 incomplete. mentioned but unclear or of experiment. experiment.
confusing.
Background information Background information Background information Background information
includes unrelated included but not sufficient provides context for giving context to goal
information and/or is to support experimental question. included.
insufficient to support goal/hypothesis.
goal/hypothesis.
Little organisation of Organised from broad to Organised from broad to Organised from broad to
information leading into specific with respect to specific with respect to the specific with respect to the
the goal/hypothesis. the topic. topic. topic.

Finishes section with brief

summary of approach.
Methods Basic Competent Proficient Advanced
(Data and Describes modelling Describes modelling Describes modelling Describes modelling
Analysis) procedure. procedure. procedure. procedure.

Provides a general Provides enough Provides enough detail for Provides enough detail for
description of the information for someone someone to repeat the someone to repeat the
experiment with little to infer how to do the experiment from the experiment from the
detail, making it difficult experiment, but details instructions. instructions.
for a reader to repeat may be unclear.
the experiment.

Inclusion of too much Inclusion of some results. Written in narrative Does not include
result. (paragraph) form. unnecessary detail - written
to an audience familiar with
the topic.

7
 Not written in narrative Not written in narrative Written in narrative
(paragraph) form (paragraph) form (paragraph) form.
Results Basic Competent Proficient Advanced
Results incompletely Results are presented in Results are presented in Results are presented in
presented in narrative narrative form. narrative form. narrative form.
form.

Tables and figures Tables and figures cited, Tables and figures are cited Tables and figures are cited
present, but not but not necessarily in appropriately and appropriately and numbered
referenced or cited correct order. numbered by the order by the order addressed in the
properly. addressed in the text. text.

Data mixed with much Data mostly free of Data is free of Apparent trends in data are
interpretation or interpretation or interpretation or identified.
conclusions. conclusions. conclusions.

Data is free of interpretation

or conclusions.
Discussion Basic Competent Proficient Advanced
Explanation loosely Results interpreted and Results interpreted and Results interpreted and
relates to results explanations offered for explanations offered for explanations offered for
collected. trends or patterns as well trends or patterns as well trends or patterns as well as
as anomalies in the data. as anomalies in the data. anomalies in the data.

Original goal/hypothesis Addresses results with Addresses results with

mentioned, but unclear respect to original respect to original
how results relate. goal/hypothesis. goal/hypothesis.

8
 Poses further questions to
continue research or address
limitations with current
project.
Conclusion Basic Competent Proficient Advanced
Results summarised. Results summarised. Results summarised. Results summarised.

Conclusion not made Conclusion relates to the Conclusion is clearly Conclusion is succinct and
with respect to goal/hypothesis but is related to the clearly answers the
goal/hypothesis. lengthy or unclear. goal/hypothesis. goal/hypothesis.

Includes summary of the Includes a succinct summary

implications of the results. of the key implications of the
results.
References Basic Competent Proficient Advanced
Few references cited References cited include References cited include all References cited include all
from text listed. most sources cited in text. sources in text but listed in sources in text, listed in a
an incorrect format as per single and consistent format.
APA.

May be listed in correct May be listed in correct Only those references cited Only references cited in the
format as per APA. format as per APA. in the text included in text included in reference
reference list list.
No references other No references other than Additional references used. References other than the
than the textbook used. the textbook used. textbook used.

Basic Competent Proficient Advanced

9
Format and Some/all of narrative All narrative written in All narrative written in past All narrative written in past
Style written in present or past tense. tense. tense.
future tense.

Frequent use of 1st Some inconsistencies Consistent use of 1st or 3rd Limited use of 1st person.
person context. between 1st and 3rd person.
person context.

Written to an audience Written to an audience Written to an audience Written to an informed

with limited knowledge with limited knowledge of with limited knowledge of audience generally familiar
of the material. the material. the topic. with the material.

Many grammar and Several grammar and Only a few grammar and Minimal to no errors in
spelling errors present. spelling errors present. spelling errors. grammar and spelling.

10