St. Paul Budget Explorer

Overview and Motivation

The St. Paul Budget Explorer was created by me and David Ma as our final project submission for CSCI 5609 - Visualization, taught by Dan Keefe in the Spring Semester, 2022, at UMN-TC.

The course focused on how to most effectively visualize data, e.g., how to represent complex data in visually understandable ways, how to generate visualizations for multivariate data sets, and how to best lay out visualizations so users can draw conclusions from them.

Our goal with the Explorer was to showcase and apply what we learned in the course to data that was relevant to us - the publicly available budget data of St. Paul.

To that end, the Explorer offers a variety of views that offer different perspectives and different insights into how St. Paul’s budget changed over several years.

Our project was awarded with the “User Achievement” award (i.e., it did the best job at providing users with means to “achieve something”, or draw insights, from the data) as one of the 3 best submissions in the class.

The application is written in Processing.

Demo Video

Part of our submission involved a demo video (narrated by me!); it gives the best sense of the data exploration abilities the project provides.

Visualization Techniques Used

Overview first, find and filter, details on demand: the initial view provided to users should be broad and give a general sense of what’s in the dataset. Options should be available to find specific aspects of the data (e.g., how much of each year’s budget was dedicated to a certain category) and filter data (e.g., just comparing two years). The most granular details need not necessarily be included in the visualization itself to prevent clutter, but should still be readily available regardless (e.g., as tooltips).

Animation: to illustrate and draw attention to change over time, animation is used. It’s important for animation to not distract from the insights you want users to be able to discern.

Wizard of Oz: what’s shown in a demonstration video need not be 100% real! It’s acceptable to set up and showcase a not-entirely-functional or fleshed-out demonstration if it is illustrative of future work or potential improvements.