Visualizing progressions for game and education design.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
final
paper
src
.gitignore
LICENSE
README.md
index.html
req.txt

README.md

Visualizing Progressions for Education and Game Design

Eric Butler, Rahul Banerjee {edbutler,piscean}@uw.edu

This project is an interactive visualization of progressions from games. It allows for two progressions from the same game to be compared.

This is useful because it lets game designers ask important questions about their game without having to playtest and re-design in a labor-intensive loop. While playtesting is critical and irreplaceable, certain insights can be gleaned by visualizing progressions in a tool like ours. For instance:

  • How does the concept ordering compare across two progressions?
  • How does the complexity change over the progression?
  • How much of the "design space" is covered by one progression, but not the other? i.e., does one progression neglect some concepts or combinations thereof?

Materials

Paper Poster

How to Run

The program is web-based, but requires a python web server backend to do the MDS computation. The primary python packages it requires are:

  • Flask
  • numpy
  • scipy
  • matplotlib
  • scikit-learn These packages are all available in pip, and can be installed with
pip install -r req.txt

The entry point is src/main.py. It will launch a webserver on localhost:5000.

So, for example, on a recent ubuntu distro, the program can be setup and run with:

sudo apt-get install python-virtualenv
virtualenv venv/
source venv/bin/activate
pip install -r req.txt
cd src/
./main.py

The open http://localhost:5000/ in a browser.

Several datasets are included in src/data/. The main page has links to a few interesting example comparisons, or datasets can be selected from a dropdown automatically generated by scanning the src/data folder.

Work Breakdown

The research process was primarily exploring several different visual encodings of our data, then integrating the promising ones into an interactive tool. We each implemented several different prototype encodings for the first half of the project, and then we jointly implemented the final interactive application, with each of us taking charge of different visualizations in the final application. For example, Rahul handled the bar chart and Eric handled brushing and linking. A substantial portion of the work was creating the various progression datafiles; we need to choose a standard format and extract/convert progression data from the games. These test encodings and conversion scripts are not included in the public submission.