In this project you and a team of other students will band together to build a non-trivial static visualization that will be rendered into the physical world in the form of printed poster. In doing this work we aim to challenge both your design and javascript skills. Once all of the projects are turned in we will have a poster session, it'll be great!
You are free to choose the topic of the visualization. It should be something that interests you and your teammates. This could be something from your research, or something you've always wanted to explore. You are welcome and encouraged to apply knowledge from other disciplines (statistics/math, machine learning, science, arts, history, etc). A debatably true adage in visualization says that an effective visualization begins with a question, you are welcome to try answer any data questions you might have.
If you aren't sure where to begin, feel free to check out the links below which contain lots of interesting data sets:
- Kaggle (a data science competition website )https://www.kaggle.com/datasets
- City of Chicago Data Portal https://data.cityofchicago.org
- Uber movement https://movement.uber.com/cities?lang=en-US
- This blog post listing datasets https://www.dataquest.io/blog/free-datasets-for-projects/
Finally, a few words of caution.
- As you start to plan your project, don't be too ambitious in your project and designs.
- Make sure you can get the data you want.
- It is very helpful to spend time sketching and exploring your data before you get to your final visualization.
- It is a common experience to make intermediate charts or diagrams on the road to producing a final graphic.
There will be three deliverables in two stages to this project
Stage 1 -> Proposal
Stage 2 -> A pdf of your visualization follow the exact dimensions we specify -> A write up of your visualization
All of these will be collected from your SVN repo. Once the proposals have been turned in your repos will be seeded with an additional scaffold where you can build your work, which you are free to use or ignore at your leisure.
In order to make sure that the project is kept in a good scope, we ask that your team prepare a document describing what you will do in the project. This project shouldn't be terribly long, a half page of text at most. You are welcome to include sketches or other diagrams that makes it clear what you plan to do. The main thing that we are looking for is that you have a reasonable plan.
You should make sure to include
- The dataset that you wish to use in your project
- How you will approach making the dataset visualizable
If you are feeling shaky feel free to chat with course staff.
We will be printing posters that are 36 by 24 (or 24 by 36 if you like) inches. It is absolutely critical that the document you prepare exactly match these proportions or we won't be able to print them. This notable comes along with a stiff penalty to your project grade. In the project-1 scaffold we list out some dimensions which have worked well in the past.
In the p1-scaffold we provide a little scaffold that'll make executing the project reasonably easy. You are welcome to use it if you like, but you are not required.
- A successful project will be aesthetically rich, well thought out, make good use of d3, present information clearly in a fashion couched in the theory of visualization we have been studying.
- A good project will be honest in it's rendering. This is clearly a complex topic in data visualization (here's looking at you, Tufte's lie factor), but in this case we mean something specific. You should not falsify or unfairly manipulate your data in order.
- A good project will show care and craft. Data visualization can be seen as a science, an art, and a craft all at once. A successful project will will show good elements in each of these categories.
- A good project uses good tools. While it is strongly encouraged that you do any number crunching/data science-y stuff directly in javascript, you are allowed to bring in any other languages you feel are necessary, though you should check in with course staff if it feels like it is diffusing the intellectual challenge of the assignment.
- You can go check out the hallway on the second floor of JCL for some inspiration
Here is the tentative grading structure for the project
- Write up: 35%
- Good javascript style (lint + eyeball), and effective d3: 30%
- Aesthetic quality: 20%
- In this meaning of aesthetic, we are not simply interested in the graphical attractiveness. We are also interested in meaningful and intriguing results. An ugly poster than contains a cool idea or piece insight about data will fare just as well in this category as a beautiful one.
- Graphic can be printed (follows the correct dimensions): 10%
- Proposal turned in on time and followed: 5%