Stream Cleaning: A Simulation to Increase Awareness & Action

Spring 2017
Role: Interaction designer
Partners: Proctor Creek Stewardship Council, Georgia Tech Public Design Workshop

Prompt: Create a simulation that will inform people about an important and complex topic

During Spring 2017, I took the course Principles of Interaction Design while also being involved in the Public Design Workshop. When it came time to decide on a final project topic, I knew I wanted to focus on the issue of stream pollution in lower-income neighborhoods. By interviewing residents, volunteers, employees and community researchers at the Proctor Creek Stewardship Council, we discovered the major problems.

Problem: Stream pollution creates health and safety hazards for residents living along it.

Most often, contamination happens in lower-income neighborhoods where many residents are unaware of trash/recycling guidelines, the local government doesn’t feel accountable for the residents, and local businesses view the area as grounds for dumping. This creates a dangerous cycle where the water gets contaminated and sewer drains get clogged, leading to flooding, the flooding leads to mold and asbestos in homes, and the next time homes flood, the mold and asbestos goes back into the creek water. The creek therefore become a health and safety hazard.

Solution: Interactors learn what they can do in the real world to help solve this problem by taking on the role of a resident in a low-income urban neighborhood along a contaminated creek.

Through the replayable simulation, they realize what types of events (with varying levels of control) occur in the neighborhood and how their choices impact the creek’s cleanliness and the community.

Design Research: Research included a comparative analysis with artifacts that were similar in subject or scope.

Simulations involving streams were rare. One was primarily an educational tool for learning about microorganisms in streams and set up like a quiz, but none covered the topic of stream contamination. Therefore, the comparative artifacts cover different topics but are similar in their goal to expose and help the interactor learn about a societal/environmental problem.

Comparative Artifact 1: Spent (
Spent simulates what it’s like to be homeless by putting the interactor in the role of a person who needs to hold a job and cover their families living expenses without going broke and losing their job. They start with $1,000 and try to make it through 30 days, having to make a monetary decision each day.

Comparative Artifact 2: Smog City 2 (
With Smog City 2, the interactor has the power to control weather conditions and emissions and must find the combination that creates least air pollution.

Comparative Artifact 3: BBC Climate Challenge (
In this simulation, the interactor is in the role of president of the European Nations. They must address climate change and remain popular enough with voters to stay in office. They have resources to put towards adopting different policies and also have to negotiate with other continents to reach their goal. The game starts in 1990 and updates every 10 years.


Design Requirements

Based on user research, comparative artifact research, and the assignment requirements, I came up with a set of design requirements for the simulation.

  • Uses Javascript and HTML5
  • Is replayable: multiple parameters, multiple distinct outcomes
  • Helps interactor understand scenario
  • Explains a complex system
  • Based on event—>choice—>outcome flow
  • Provides real world information through events, choices and outcomes
  • Puts interactor in role with amount of control reflective of real world scenario
  • Displays transparent relationship between variables and provides immediate feedback
  • Uses transparent visuals for parameters and context

Iterations & Evaluations

The Stream Cleaning simulation began as a flow diagram in the context of a specific urban creek with the interactor in the role of a resident in a low-income neighborhood along the creek.


The events would be the same for each interactor, but their choices would affect the three variables of water purity (initiallly water contamination), local government/business sense of accountability, and resident morale in the neighborhood. The end state would be determined by the measure of the variables at the last stage of the simulation with 3 potential outcomes: the neighborhood thriving with a clean creek, the creek remaining contaminated, or the creek causing so much damage through flooding that the residents are forced out of their homes.


Updates based on mockup feedback:

  • Presented variables at beginning
  • Represented variables with icons
  • Kept variables in one location throughout simulation and on consistent scale
  • Added a parameter for overall time/energy of resident
  • Decreased text—decided what to have on screen and what to have as optional additional information/context

With the next round of testing, the goal became to make the final design more replayable, to make the relationship between variables more transparent, and to show as little text as possible at any given point throughout the simulation. Therefore, I made the following updates:

  • Differentiating between categories/relationship between variables with color (separate water purity (effect) from sense of accountability and resident morale (cause))
  • Adding a fourth potential ending where the simulation would end early and go to the “disrepair” end state if any of the variables hit “0” before time was up to make the simulation more replayable
  • Bolding phrases so that interactors could more easily skim text
  • Removing links in navigation to external sites so that interactors instead must complete the simulation before having the option to leave the simulation for more information
  • Adding introductory interfaces where the interactor could decide whether or not they wanted to attempt the simulation

Future Plans

Based on testing and comparative artifacts, the design of the game could potentially be enhanced with a few additional features.

Time-step: While the “30-day” length of Spent seems to make the simulation less replayable, Stream Cleaning could have more events with decision points as well as random uncontrollable events.
Avatars: Considering that Stream Cleaning takes more than one role into account, the future design could give the interactor the option to experience the issue in the role of government or a business instead of only having the option to be a resident.
Visual: One of the most powerful parts of the stimulation is making the interactor aware of the problem by showing images of the negative impact. This could be heightened with a visualization of the water’s purity rather than or in addition to the progress bar.