COSC3000 S1 2025
Visualisation project
DUE 15TH APRIL, 2025 OR 16TH MAY, 2025: These are the two dates by which you will submit the visualisation project and the computer graphics project. You can choose which of the two projects you will submit by the earlier date (15/4) and which by the later date (16/5).
Task
Data from a UQ PhD project is provided. The data are from play sessions, each with a child with autism playing with a typically-developing peer play partner. Some of the peer play partners were trained, and some untrained (control group), and the level of social interaction and cognitive play were measured to determine the impact of the training of the peers.
The thesis (serene_choi_thesis.pdf), and a book chapter (serene_choi_nova_chapter.pdf) with a more compact version of the results and discussion and simplified figures are provided. The data is plotted in chapter 7 (pp 145-194) of the thesis.
Analyse and visualise the data, to present the results appropriately for a book chapter or paper like the provided book chapter (for print or online reading) or a conference or seminar presentation of those results (for projection)
You should make the following improvements:
• Use colour.
• Make it visually obvious whether the changes are statistically significant at the 0.05 level.
• The results showing the overall social and cognitive interaction in the book chapter and thesis don’t show the individual session results (in the thesis, the individual session results are shown for the specific types of social and cognitive behavior. (see section 7.4 in the thesis). Make test visualisations to see if the individual session data can be shown easily for the overall social and cognitive behavior.
• The weights used to convert the data to an overall cognitive and social score for each session (see table 5.6 (pg 120) in the thesis, or table 8 in the book chapter; note that the “non-play” and “stereotype” weights are the wrong way around – they should be 0 and - 1 rather than -1 and 0 ) could be chosen to have different values. Make a visualization showing the effect of reasonable variation in these weights on the results.
Your project report should show your visualisations, and give explanations of your design and reasoning.
The data
While there are relatively few play sessions, you can assume that the means are normally distributed and therefore use the standard errors in means to determine confidence intervals. If you prefer, you can use appropriate small-sample methods. A Matlab file and a python file to read and process the data is provided.
The names of the children involved in the experiment have been anonymised in the data files. Children with names beginning with A-E are the children with autism, U-Z the trained peer play partners, and L-P the untrained peer play partners. The Matlab read_dyad_data function can be used to read the data for all pairs:
t1 = read_dyad_data('albert','ulrich');
t2 = read_dyad_data('barry','viola');
t3 = read_dyad_data('chris','wendy');
t5 = read_dyad_data('dana','xavier');
warning off
t4a = read_dyad_data('ellie','yoshi');
t4b = read_dyad_data('ellie','zara');
warning on
u1 = read_dyad_data('albert','lydia');
u2 = read_dyad_data('barry','mario');
u3 = read_dyad_data('chris','nellie');
u5 = read_dyad_data('dana','oscar');
u4 = read_dyad_data('ellie','peter');
Note that the data for Ellie was omitted from the original figures, since the original trained peer withdrew from the project. The data is stored in the assigned variables as structures. The main data of interest in the variable t1 above will be in
t1.score
t1.score_error
t1.session_score
etc., where “score” is the average score over sessions, “score_error” is the standard error in that average, and “session_score” is the scores for the individual sessions. There are 8 groups in each (cognitive/social interaction, indoor/outdoor, pre-training/post-training), which you can access like:
t1.score.cog_in_pre
etc. If you are using Matlab but haven’t used structures before, you might like to try the command fieldnames, e.g.,
fieldnames(t1)
Feel free to ask questions about the data and how to read it in the labs and on the discussion board.
Marking
The project is worth 0-8 marks:
0: fail
1-3: 4 and low 5 – major errors or omissions in the required tasks
4: high 5 – required tasks are all completed, but there is no evidence of further achievement or deeper understanding
5-8: 6 and 7 – required tasks are all completed, and work shows frequent or consistent evidence of advanced skills/understanding.
You must pass the project (i.e., obtain 1 mark or more) to pass the course.