Assignment 2: Visual Analytics Practice
As outlined in grading and rules, to complete the course and take the credits, students should complete two assignments. This page describes the second (practical) assignment. The first (theory) assignment is described here.
Besides testing the student's abilities to evaluate and improve a visual design, the course aims to test the ability of actually constructing a visualization that answers specific questions. This is the aim of the current practical assignment. You will use a real-world (large, unstructured, time-dependent, multidimensional, hybrid) dataset, and you will have to choose, construct, and use several visualizations to answer a set of specific questions on the dataset. In addition, this assignment will train (and evaluate) your abilities to use Tableau, a state-of-the-art visual analytics framework.
The assignment has the following steps. A detailed description is available in this document.
Step 1: Install Tableau
Visit the Tableau site to install a working copy of the Tableau visual analytics tool on your computer. Note that several versions are available. For this assignment, you can use either the Tableau Public or the Tableau Desktop versions.
Step 2: Get the data
Visit the site indicated in the assignment document, and download the indicated dataset. Note that this is a quite large dataset (over one million records, several hundred MB's in size). The downloaded dataset comes as a CSV (comma separated values) text file. Verify that the dataset is properly downloaded by opening it in Tableau and browsing the records. Compare the values you see with those displayed in the online browser available on the download site.
Note: For back-up reasons, an offline copy of the above dataset is available here. Use this only if you have difficulties in getting the live dataset from the online website, and only after having discussed such issues with the teacher!
Step 3: Perform the assignment tasks
Once the data is properly loaded in Tableau, you will have to complete five tasks, each consisting of several sub-tasks (questions). These questions are formulated on purpose at a high level, as they try to mimic the types of questions one encounters in practice when analyzing big data. For each question, design one or more visualizations in Tableau that answers the question. Use your designed visualizations to obtain answers to the questions. Describe, in detail, how you have designed the visualizations (and why you have preferred the respective designs); argue on their suitability but also on their limitations; if other visualizations, not present in Tableau, would be more suitable/effective for answering the questions, describe this (and why); finally, describe your findings (answers to the assignment's questions).
The final deliverable for this assignment consists of
- a detailed report (PDF) describing all steps 1..5 of your assignment. Support the description of your results (and challenges found) by snapshots from your actual visualization, wherever this is appropriate;
- a binary deliverable of the visualizations you created in Tableau to solve this assignment; the easiest for this is to share your visualizations publicly using Tableau's Online cloud server. Please consult the lecture slides to see how this is done.