Data Science is a rapidly expanding field in industry and many leading universities in the USA and the UK are starting data science degrees and units. Monash FIT trialed a data science unit for the undergraduate business IT course, FIT3152, in 2013, which is continuing, mostly focusing on data analysis. This Modelling for Data Analysis unit is a core elective in the full Master of Data Science to start in 2016, and a core unit in the Graduate Diploma in Data Science (Monash Online) starting 2015.

23/2/2015: Amended wording in following fields for administrative clarification: Reasons For Introduction; Role Relationship Relevance; Teaching/Mode; Assessment/Summary; Workload/Contact Hours; Location Of Offering

27/2/2015: Amended Workload Requirements following discussion with Chair of GPC to clarify the workload required for the various teaching modes.

07/08/2015: Admin - submission of changes to learning outcomes and synopsis made by CE in March. Also numbered learning outcomes.

17/09/2015: Admin - CE modified learning outcomes on 11/08/15.

18/4/2016: Amended assessment to give on-campus version exams, instead of using Respondus lockdown browser for equivalent in-semester assessment. (K Korb)

31/8/2016: Amended outcomes based on Matt Bangerter's general meeting and pre-reqs based on Trudi Robinson's general meeting. (W. Buntine)

01/09/2016: Admin - numbered learning outcomes.

07/09/2016: Admin - Updated learning outcomes as per GPC 4/16 (Item 7.1).

06/01/2017: Admin - Lecturer updated assessment to align with University policy. Effective Semester 1, 2017.

20/01/2018: Wray Buntine - decrease EOS exam to 50% since only 4 outcomes tested

06/02/2018; Wray Buntine - add ETC5252 as prohib. since its now prereq for FIT5149

29/05/18: Wray Buntine -- standardised assessment entry, doesn't affect exams but clarifies things, so useful but not necessary for S2 2018

24/9/2019: Admin - adding 10 minutes reading time to the overall exam duration as per University requirements.

3/10/2019: Unit name changed to Statistical data modelling, effective semester 1, 2020. With the introduction of the Master of Artificial Intelligence and changes made to the Master of Data Science, it was agreed that the name of this unit would be changed and are already reflected in the 2020 courses.

18/10/2019: Updating the prerequisites. As FIT9133 is no longer offered, adding FIT9136 for commencing students. Effective semester 1, 2020.

11/11/2019: In addition to the prereq update entered 18/10/2019, also adding a prereq waiver clause for students studying C6007 Master of AI. Entry into C6007 requires cognate studies with relevant maths and programming; and as such, students meet the FIT5197 enrolment requirements. Effective 2020.

18/09/2020 - Admin: Update to include new assessment and teaching approach fields as per Handbook requirements.

21/01/2021 - New 6th edition of text book is now available and has been updated here.

22/01/2021 - Reduced number of assignments from 3 to 2 to reduce stress and workload on students in 0n-Campus offering. Aligned learning objectives with the updated edition of the textbook. Included a recommended reading called 'Learning R' since many students are not strong in R programming and need a support text book. Updated technological requirements to say RStudio is needed for tutorials and Jupyter Notebooks with R is needed for assignments as this is how we have been delivering the unit.

22/02/2021 - Learning outcome 2 has been modified to read as follows at the request of the faculty: Construct mathematical models and apply them in different use cases

2/3/2021 - Aligned ULOs of assessments for online and on-campus versions

7/7/2021 - Changed from laboratories to tutorials in workload because we have always been treating them as tutorials and tutors have been charging them as tutorials but Wray never initially got around to changing monatar. Assignment 1 converted to mid-semester test to avoid plagiarism. Removed references to Jupyterhub in techical requirements because it is not used anymore.

This unit will be a core elective unit for the Master of Data Science starting 2016 and is a prerequisite for the planned new Data Analysis unit FIT5201. It is also a core unit the Graduate Diploma in Data Science (Monash Online) starting 2015. It is also a graduate variant of the undergraduate FIT2086.

There are no existing masters units offering the same material as this unit.

The unit covers general statistical theory behind data science.

Upon successful completion of this unit, it is expected that students will be able to:

1. Perform data analysis with descriptive and inferential statistics on given datasets
2. Construct mathematical models and apply them in different use cases
3. Produce models for inferential/predictive statistical analysis
4. Perform fundamental random sampling, simulation and hypothesis testing for required scenarios
5. Implement mathematical models through programming and scripting
6. Interpret results for a variety of models

020399

This unit explores the statistical modelling foundations that underlie the analytic aspects of Data Science. Motivated by case studies and working through examples, this unit covers the mathematical and statistical basis with an emphasis on using the techniques in practice. It introduces data collection, sampling and quality. It considers analytic tasks such as statistical hypothesis testing and exploratory versus confirmatory analysis. It presents basic probability distributions, random number generation and simulation as well as estimation methods and effects such as maximum likelihood estimators, Monte Carlo estimators, Bayes theorem, bias versus variance and cross validation. Basic information theory and dependence models such as regression and log-linear models are also presented, as well as the role of general modelling such as inference and decision making, and predictive models.

Required Reading

Ross, S.M. (2020) Introduction to Probability and Statistics for Engineers and Scientists, 6th ed. Academic Press.

Recommended Reading

Cotton, R. (2013) Learning R, O'Reilly Media Inc.

Technological requirements

Students may use Windows, Linux or Mac environments for this subject.

On-campus: R and RStudio must be used for tutorials and Jupyter Notebook capable of running R must be used for programming assignments. Non-programmable calculators can also be used for tutorials and exams. Rulers for drawing graphs are also allowed in the exam.

Monash Online: Jupyter Notebook capable of running R or RStudio must be used for programming assignments.

On-campus, Monash Online

Monash Online in-semester assessment: 100%.

1. Assignment 1 - Basic Data Modelling - 20% - ULO: 1, 2, 4, 5
2. Assignment 2 - Exploratory Data Analysis, Statistical Inference, Modelling and Prediction - 30% - ULO: 1, 2, 3, 4, 5, 6
3. Final Quiz - 50% - ULO: 2, 3, 4, 6

On-campus examination (2 hours and 10 minutes): 50%; in-semester assessment: 50%.

1. Mid-semester exam - 20% - ULO: 1, 2, 4, 5
2. Assignment 2 - 30% - ULO: 1, 2, 3, 4, 5, 6
3. Examination - 50% - ULO: 2, 3, 4, 6

Minimum total expected workload equals 144 hours per semester comprising:

1. Contact hours for on-campus students:
   • Two hours/week lectures.
   • Two hours/week tutorials.
2. Contact hours for Monash Online students:
   • Two hours/week online group sessions.
   • Online students generally do not attend lecture and tutorial sessions, however should plan to spend equivalent time working through resources and participating in discussions.
3. Additional requirements:
   • A minimum of 8 hours per week of personal study (22 hours per week for Monash online students) for completing tutorial activities, assignments, private study and revision, and for online students, participating in discussions.

FIT

(FIT9131 or FIT9133 or FIT9136) and MAT9004 or equivalent; or entry into C6007.