CS 8395-03 - Visual Analytics & Machine Learning




Project

In the second half of the semester, you will form teams of two to work on a project.

Project Types

The type of project you choose should come from three of the main research topics that we will cover in the course: Mixed-Initiative Visual Exploration, Visual Analytics for Understanding Models, or Visual Analytics for Training Models.

The first two topics are most suited for development with Observable notebooks, and should be natural extensions from the programming assignments. Visual analytics for training models, however, may be more suitable for a client/server model. If you would like to go this route, then please contact me in advance. Additionally, if a project that you would like to do does not quite fit in the above topics, then again, please contact me in advance.

Team Formation and Project Proposal

You will first form a two-person team as part of the project. If you wish to work in a team by yourself, or a team of three, then please let me know.

You will then formulate a project proposal. The proposal should include the following:

Unlike the assignments, you will share your notebook with me. This does not make your notebook visible to the public, rather, anyone with the link will be able to access it. If you have any problems with this please let me know.

Additionally, each team will present their proposal to the class. Please prepare a 10-minute presentation.

Project Prototype

The project prototype should be comprised of three parts:

Data Wrangling

You should be able to complete any data cleaning/wrangling/munging at this stage. My recommendation is to take data sources that you’ve collected, or data generated via machine learning models, and write them out as JSON files for further processing on the visualization side.

Exploratory Data Analysis

You should quickly prototype some visualizations for exploring your data. I strongly recommend using Vega-Lite for this purpose, as it simple to use and integrates with Observable notebooks, while still providing a decent-sized design space to consider. In your notebook, you should present these Vega-Lite visualizations alongside discussions written in Markdown, detailing any findings regarding your data.

Design Prototype

You should have a working prototype of your project in place. The basics should be here: spatial organization, visual encodings, some (but not all) interactions, and integration with machine learning models, as appropriate. It does not need to be polished yet, but rather, it should be clear where the project is headed.

Project Presentation

Each team will have 10 minutes to present their project. Your presentation should include the following:

Final Submission

Your final submission should contain your final visualization design, as well as a discussion about your project:

Project Assessment

As everyone’s projects will be different, there is no detailed rubric for grading. Nevertheless, the projects will be assessed based on their intended contributions. In general, this should consist of the following components:

However, if the main contributions of your project deviate from the above, then the assessment will be adjusted accordingly.

Project Schedule