1 of 12

FiTrack

By Jimmy Ding and Chong Wong

Declutter your mind

2 of 12

Why this matters?

According to [1], Decision Fatigue is what drains you physically and mentally throughout the day as you make hard decisions. If writing down your agenda can help clear your subconscious mind from clutter and increase productivity, we wanted to create an app to help the end users decrease the amount of worrying they have in life. Our app relies on the fact that you let the machine decide when you exercise based on machine learning in order to add variety [3] to improve your training regime (by decreasing boredom and plateaus), and declutter your mind. This is especially relevant as coronavirus has made us more sedentary sitting in front of a computer - harming our posture and health.

[1] - https://www.psychologytoday.com/files/attachments/584/decision200602-15vohs.pdf

[2] - https://pubmed.ncbi.nlm.nih.gov/21688924/

[3] - https://www.acefitness.org/education-and-resources/lifestyle/blog/1210/why-is-it-important-to-vary-my-workout-routines/

3 of 12

How we do it?

React Native

Expo

Machine Learning

4 of 12

Live Demo!

5 of 12

Setting up the profile

  • Height
  • Weight
  • Age
  • Past activity

6 of 12

Factors For Planning The Schedule

White and black listed times entered by the user

Historical data of whether the user started the workout

The user’s current location and geofencing data entered by the user

Holidays and weekends

The user’s activity throughout the week

Distance towards the user’s goal

7 of 12

White and black listed times

White List

Temp White List

Less White List

Black List

Temp Black List

Less Black List

Significantly increases the chance of sessions being started

Same as white list, but gets deleted after the time passes

Slightly increases the chance of sessions being started

Disallows sessions from being started

Same as black list, but gets deleted after the time passes (use case: traveling)

Decreases the chance of sessions being started

8 of 12

Historical data

Factors:

Time in week - for non-uniform schedules such as college class schedules

Time in month - for monthly events such as periods

Time in year - for major holidays such as thanksgiving

9 of 12

Geofencing

Factors:

Distance from home - To detect events such as travelling

Distance from nearby gyms - To notify user when they are close to a gym

Black List Areas - To avoid areas such as workplaces or schools

White List Areas - To increase the chance in areas such as home

10 of 12

User’s activity

Factors:

When is the app opened - detect range of leisure time for user

Amount of temporary lists added - detect whether a user is busy for the week

Amount of manual entries added - detect whether a user is free for the week

11 of 12

Distance Towards User’s Goal

User is greatly underperforming - slightly lower the goal to adapt to the user

User is slightly underperforming - increase the frequency in order to reach the goal

User is slightly over performing - decrease the frequency to give the user rest

User is greatly over performing - slightly raise the goal to adapt to the user

12 of 12

Future plans

  • More advanced algorithms and techniques
    • Anonymously collect user data to develop a deep learning algorithm
  • Calorie estimator using gps and step tracking
  • NLP to import the calendar schedule