A2: Play with Rec Systems

You are going to have experience with a few visualization design recommendation techniques. These are listed as the first three “optional” readings under Design Recommendation.

Due: Friday 10/1 11:59PM EST..

Draco

Draco is a system that models visualization design as a constraint problem based on Answer Set Programming. It is meant to be used as a recommendation engine within an exploration or recommendation tool such as Voyager. In this assignment, you will try your hand at writing Draco specifications, XYZ, and comment on the experience.

Go to this Observable notebook and

1. Create an account and fork the notebook
2. implement the createDracoProgram() function.
3. Click “publish” in your notebook and remember the URL

Voyager

You will use Voyager for data exploration and report your thoughts

• Go to the voyager website and load the “Birdstrikes” dataset. This is a dataset from the Federal Aviation Agency which collects instances of birds that strike planes, as self-reported by airlines, pilots, and airports.
• Use the tool to identify two interesting facts from the dataset.
• Reflect on the ways that Voyager assisted your analysis, as well as ways where it was not useful or negatively impacted your analysis process.
• Optional: Submit the vega-lite specs of the views you generated in the assignment form.

The following are instructions to generate the vega-lite specs.

1. Each view has a small bookmark icon, if you click it twice it will ask if you want to remove or keep the book mark.
2. On the top left, there is a “Bookmarks (x)” link that will list your bookmarks. Click on it.
3. Click the “Export” on the upper left of the overlay to download the vega-lite specs of the bookmarked views.

Lux

Lux is an intent-based visualization recommendation system, where you specify attributes of interest and it recommends visualizations based on them. You will load the example Jupyter Notebook and go through the tutorial. At the end of the notebook use their intent and visualization mechanisms to identify a new interesting fact in the COVID data (or another dataset).

• Lux github
• Live Jupyter Notebook

Submission Instructions

Submission

• Use this submission form and answer the reflection questions.
• Start thinking about projects you want to propose!