Up One Up
I love going for a run and ambition has always driven me to do better and faster. There are some really good apps to help you track your progress and even offer different training plans to help you reach the next level.
Wouldn't it be interesting to specifically explore how AI and machine learning could tackle your training? An application with primarily the focus to push you to your limits? Combine this with the power of wearables such as the Apple Watch? This was the direction I wanted to explore, with Up One Up as the result.
While looking at existing apps, a frustration for me is the limitations of existing training plans to become a better runner. Based on your average time per kilometer, your heartbeats per minute, age and community-wide data you can produce a personalised training plan.
Specifically the notion of slightly improving your workout with very specific goals such as:
- Run a minute longer than previously (time-based goals versus distance)
- Increase run speed and heart rate (anaerobic exercise)
- Decrease run speed and heart rate (endurance training)
- Contextual goals to improve your training (use Apple Health data as reminder to sleep enough for muscle recovery for example)
The above goals, which can vary during your training sessions greatly increase motivation by using smaller, achievable goals versus typical distance or purely time-based training plans.
Beyond a goal-based application, I believe in the power of useful and beautiful data tracking. Presenting data in nice charts can help as extra motivational factor beyond the typical gamification.
Beyond being goal-based, the application should provide a visual dashboard containing the user's progress.
ON RELEVANT COACHInG
Every athlete is slightly different, hence there doesn't exist a perfect training plan. What's great about applying machine learning is that you can use data provided by smart devices (heartbeats per minute, distance, age) to learn the personal limits of the athlete. Based on that, you can continue tweaking the plan until it's optimal for the user's profile, comparing the data with the community-wide datapool.
I figured it would be interesting to add a specific example, to explain how machine learning and the use of data can affect a runner's performance.
Your VO2 Max rate is a good indication how much your body is suffering. In essence, it's the indication of the maximum effort your body is capable of doing. The guiding factor here is your heart rate (hence wearables are crucial for an app such as this one to work). Depending on your age, your maximum heart rate decreases.
Personally, my max. heart rate is around 195. When I'm running and have trouble hitting a certain distance goal while having a consistent heart rate of 190, that's a good indication that the training plan is too ambitious and should be altered. This is data which the system can learn.
The input to make this work is combining the data of the running community which uses the app. This can provide averages of the collected data to build up a training plan. Now, the training plan based on averages can be altered based on data of your personal performance after a number of runs.
Furthermore, even if the training plan was too ambitious, that particular run would still be excellent considering I was running at my personal limits. The app in that case, should still provide positive reinforcement and congratulate the runner.
There are more external factors which affect your performance. Think for example of the weather (too warm) and the location of your run (air quality), these are more data points which can be used to correctly measure your performance and push you to do better.
To make the app succeed, I believe in making it as simple and accessible as possible.
Many running apps offer a whole range of metric tracking visible during and after your run. If you're a casual runner this can overwhelm you as you're not quite sure which metric to pay attention to.
Neither, the best course of action is presented well.
Hence, I believed working out a simple information architecture would be key to make the notion of using data to provide a customised training plan succeed.
In summary, the app would contain the following flows as a minimum viable product:
- Training - Start running with a goal presented to you
- Progress - View your improvements as a runner plotted over time
- Profile - Manage your profile settings
For the look and feel of the application, I believe in doing something energetic and exciting. To provide a better contrast with many apps which today include a lot of simple color schemes and pursue minimal design, I believe a design exploration which is less traditional could be an interesting design style to pursue.
An energetic design combined with the power of data, that's Up One Up. Keeping the app simple, by focusing on either your training or your progress keeps the product focused so that the app helps with it's objective: helping you become a better runner.
Would you like to get in touch? Feel free to send an email.