This assignment must be completed in a group of minimum 3 students and maximum 4 students.
This assignment is a prelude to the third assignment. It aims at providing you with an authentic experience in carrying a simple data science project that covers all essential stages in a data science lifecycle. Since most professional science projects are performed by teams, you are therefore required to complete this assignment in a team.
In short, you are required to complete the following tasks:
- Pitch a public, open dataset of your choice.
- Pitch 3 or 4 initial hypotheses to be pursued later in Assignment 3.
- Profile the data using descriptive and/or inferential statistics techniques (which also requires that you demonstrate proficient data wrangling skills).
- Present items 1, 2, and 3 above via a recorded presentation.
- Your tasks are open-ended tasks, similar to most real data science projects. This means no two teams are likely to go to the same direction and produce similar results. You will find that your group will become experts in interpreting your own data and answering your own problems. Comparing performance across teams may not be meaningful and your team will be assessed solely against the rubric.
Your Python code base must be available on your Github repo. The extent of the group’s collaboration and individual contribution will be evaluated solely based on Github.
- Select an open (publicly available) data – data that can be freely downloaded, preferably with an open license, allowing you to share the data freely. Choosing non-public data is not advisable as your instructor may be restricted from accessing the data.
- Choose data in the domain for which team member(s) has some background.
- Formulate open-ended hypotheses.
- Carry out fresh data and/or analysis.
- Where possible, choose a dataset and formulate problems pertaining to practical Australian contexts.
Choose only 1 dataset.
It is fine to choose a dataset that has been analysed by others outside of the university. This is the natural consequence of selecting open data. However, you should either show that the analysis and exploration you plan has not been done before, or show that there is no code already available to do the analysis you intend. Your instructor is likely to view highly any original investigation.
Sources of open datasets include but are not limited to:
Group work activities must be visible on Github Classroom.
The instructor will send an invitation to all students to join Github Classroom after all groups are formed. To accept this invitation, every student must have a free Github account. If you do not already have it, please sign up. This is compulsory.
You should refer to the detailed marking rubric that appears on the side panel of this window.
Please submit the following via Learnline by latest 11.59pm on the due date:
Academic integrity and assessment irregularities
Academic integrity is a core value at CDU and must be upheld at all times when completing this assignment. You must not plagirise the work of others. Please be referred to the Students – Breach of Academic Integrity Procedures.
Other assessment irregularities are governed by CDU’s Higher Education Assessment Procedures.
Tips and example
Broadly speaking, your instructor is looking for evidence of your demonstrative competency in the following key data science skills implemented in Python: (1) hypothesis formulation, (2) exploratory data analytics, (3) data wrangling skills, and (4) data visualisations.
When pitching your dataset, consider addressing the following concerns:
- source of data
- accesssibility of data
- validity of data
- why the dataset matters (in practical or academic terms)
- domain knowledge
- relevance to you
- In profiling the data, consider addressing the following concerns:
- data types
- shape of data
- The last task is to pitch 3 to 4 initial hypotheses. Consider addressing the following concerns:
- what might the data tells us
- what would you like to explore first based on your initial data profiling
- what would you like to predict
- what existing assumption you want to test previous finding
- what new idea you want to test