A good goal

Ask any gamer, and they’ll tell you how important goals are. So much so that it decides whether gamers chooses to play that game or not. As Jesse Schell tells in his book The Art of Game Design, a good goal must be concrete, achievable and rewarding.

Machine Learning

From our research, one clear concrete goal for everyone was to make roads safer. Good goals must also be achievable, so we built a machine learning model to automatically detect violations (and hence solve the problem of violations being overlooked). In the interest of time our professors asked us to implement the “no helmet violation” as a first cut of the minimum viable product.

My team and I developed an efficient implementation of a neural network to automatically detect violations. We used YOLOv2 algorithm (You Only Look Once) which is a unique convolutional neural network (CNN) for doing object detection in real time. The algorithm applies a single neural network to the full image, and then divides the image into regions and predicts bounding boxes and probabilities for each region. Predicted probabilities are assigned to the bounded box in the form of weights. This achieves high accuracy while also being able to run in real-time.

With a robust machine learning model in place, we made much it easier for users to get clear and immediate feedback by detecting violations in real time thus rewarding the user with assured violation detection.

Safety and privacy

Another key ask was about privacy and safety of the users using app. There were two usecases we had to design for

1. 1.

   Some users for obvious reasons, did not want to reveal their identity when a violation is submitted.

2. 2.

   If the user spots a violation while in motion in a vehicle, they cannot use the application as a driver which is both a distraction and a violation in its itself.

For the first usecase we came up with the idea of having users enter aunique pseudonymas an identifier. Their real identity is encrypted and will never be visible on any submitted violation.

To solve the second usecase, we introduced a feature that diables the ‘Book Violation’ CTA when any movement is detected by combination of the phone’s gyroscope and GPS based location. It will then trigger a quick challenge test to prove that the user is not the driver. The questions in the test is uniquely designed to test the attention of the user. Successful completion of the test will prove that the user is not distracted from focusing on the road. But if there is no constant motion detected (ie. not in vehicle), the user is directed straight to the camera without the test.

Challenges motivate users

Reaching goals is typically accompanied by challenges. At Check we built in a leaderboard to track top contributors to a better society. Motivating users to move up the leaderboard with pleasant surprises like badges and vouchers indulges them in playful competition and results in users contributing positively to build a safe environment

Strong emotions

The best games create strong emotions. When designing for games we aim to tap into nuanced explicit and implicit emotions. The following framework of the Emotion Wheel from Junto Institute explains more clearly the subtleties we have to work with. For this application our users care deeply about for the society at large which gives them a feeling joy. Our design has to be for ‘Optimism’ and ‘Hopefulness’ because our users are looking to improve their society. Our designs must also be for ‘Pride’; when a user does a service for the society which rightly gives a sense of accomplishment.

In our application, we designed a rewards section that displays the badges and vouchers won by the user for being good citizens of the society. We carefully craft messages to convey a sense of pride and accomplishment.

A holistic experience

The three previous factors are vital components to our game design approach, but the one that stitches all of it together to form a holistic experience, is the flow. Designing a flow must bring intensely and focused concentration to the task. It must be so absorbing that we don’t get distracted about the next or the previous steps. The flow must be so rewarding that the task becomes intrinsically motivating.

A vital condition to maintain good flow is ‘knowing what to do next and know how to do it’. So to help users familiarize themselves with how the app works we will show them a quick onboarding screen explaining the flow in four easy steps.

The application was mindfully designed to be extremely lightweight with just three sections: a profile page; rewards; and settings. This makes navigation through the application incredibly easy with minimum effort.