The aim of the assessment is for you to demonstrate the skills required to produce an effective data visualization sketch and evidence that you can identify and apply good practice in visualization design. You should aim to show you can abstract a data visualization task and validate your visualization design. Specifically, on successful completion of the assessment you will have:
identified (a) research question(s) about a real dataset that can be answered with data visualization;
built a working data visualization that represents a real dataset;
applied good practice in the visual encoding and interaction in your design;
provided a validation (justification) of your design;
provided some insight into your research question as a result of your visualization;
demonstrated a critical understanding of the limitations of your visualization (postgraduate students only).
Note the instructions below refer to producing data visualization using Processing (the standard 'Java' flavour, the Python flavour or the JavaScript version p5.js). If anyone would like to use an alternative technology such as D3 you should contact me first to approve the technology. Regardless of the technology you use, the assessment outcomes and marking scheme below will apply to all submissions.
The single assessed piece of coursework for this module involves you choosing one or more data sources and creating a sketch in Processing to visualize the data. You will also provide a design justification for your visualization validating the design decisions you made in creating your visualization.
You may choose any dataset from any source and you are encouraged to look widely for possible sources. Some examples of places to look for inspiration include the Guardian Data Blog, the Financial Times Chart Doctor and the Office for National Statistics (ONS) There are many hundreds of datasets available so you should spend some time carefully considering the task and data that interest you. The following guidance may help:
Choose a dataset that is sufficiently complex that it will benefit from exploration though interactive data visualization (e.g. a table of 10 numbers would probably not need anything more than a simple Excel chart)
Choose a dataset which will help you answer a motivating research question. Examples of research questions could be What are the patterns of criminal behaviour and do they show changes over time? or Is there any structure to the text in this collection of documents that suggest they have something in common?.
Avoid choosing datasets that are very difficult to parse either because they are too large, too complex or are not in a form amenable to parsing in Processing.
You may choose to link more than one dataset if that helps you answer your motivating research question. There is no minimum or maximum number of datasets you should chose.
You should create your Processing sketch in much the same way as you have been for the non-assessed data challenges. You will not be assessed on the quality of your code, but rather the design and effectiveness of the resulting visualization. You may use other libraries or PDEs in your sketch (e.g. giCentreUtils, geoMap, ControlP5 etc.) if you wish, but you must credit any other code that you have not written yourself with clear comments in your code.
If you find that you cannot implement everything you wish to in your design, your submission may include a description and/or mockup of what you intended. If you do so, your submission should make clear what is implemented and what is a mockup.
You must use the file courseworkTemplate.doc as the basis for your justification document.
This outlines the structure of your document which must be no longer than 5 pages (postgraduates) or 4 pages (undergraduates), excluding a separate page of references.
Do not change the formatting or margins in order to fit in more text (but you may remove the guidance text included in the template, leaving the headings).
You should save a local copy of the document with the name Validation_MySurname_.doc (substituting MySurname with you own).
Please ensure you use this naming convention.
You should submit a single .zip file file (do not use other archive files such as .rar) containing the following to the Coursework Submission Area on Moodle no later than Sunday 30th April, 2017, 5:00pm:
Your working Processing sketch including all relevant PDE files, the data directory containing the data you are visualizing and any other relevant files, such as font (.vlw) files, additional libraries required to allow your sketch to run (if you use any libraries from the giCentre such as giCentreUtils or geoMap there is no need to include these with your submission).
A PDF document no longer than 4+1 pages (undergraduates) or 5+1 pages (postgraduates) using the template given above outlining your design validation and, for postgraduate students only, critical evaluation of the visualization. The '+1' refers to a single page containing any references you have used to support your design evaluation. That page should contain references only, and must not be used for any other content.
Late submissions will not be marked, so you are very strongly advised to plan submitting your work well in advance of the deadline. Any material in the justification document beyond the page limit will not be marked.
You are encouraged to support each other in terms of general approaches to data visualization and the use of Processing, but you must not discuss or share details of your own coursework with any other students - this is an individual piece of assessment.
You will receive written feedback on your work within 4 weeks of the coursework deadline.
In marking your work I will be looking for evidence that you have met the aims and outcomes at the top of this page. Specifically, you will get credit for the following:
identifying (a) research question(s) that has been answered effectively through data visualization taking advantage of both the 'human in the loop' and 'computer in the loop'.
building a working data visualization sketch that provides some insight into the data you are representing and the research questions(s) you are answering
demonstrating that you have considered good practice informed by literature in the design of your data visualization
reflection demonstrating your critical understanding of the limitations of the visualization process with respect to your own design (postgraduate students only)
A submission will pass if you have demonstrated a design for a data visualization that answers your research question(s) in a way that would not have otherwise been immediately obvious by looking at your chosen data directly or could have been shown with some simple charts or calculations in spreadsheet such as Excel. By 'answering your question' I mean that you have generated some new understanding as a direct result of your visualization that relates to the question you have set yourself. Your validation must show some evidence you have considered good practice in the visual encoding and interaction design of your visualization.
For a submission to get marks in the distinction / first class range (70% or more), it should at least do everything as above, but also be able to draw on good practice in visualization design from literature mentioned in the lecture notes or from other external sources. It should demonstrate a clear link between your task abstraction and your validation. It should demonstrate some sophistication in its design and implementation that clearly goes beyond the examples provided in the lecture materials. It should reveal complex patterns in the data in a manner that is easy to interpret and that directly help in answering your research questions. You should be able to demonstrate how your own design has assisted in answering those questions. Postgraduate students seeking a mark in the distinction range should also be able to demonstrate an awareness of the limitations of the visualization process and relate that to your own design, research question and data insights.
Under no circumstances must you make direct use of any other student’s coursework. Any submissions that appear similar to those submitted by others will be investigated for possible academic misconduct.