Subtle.

CODE

BrainStation Capstone Project

a machine-learning application that analyzes poses in photographs to provide a summary of repetitive gestures

Front-End

Apollo

ml5.js

p5.js

React.js

Tensorflow.js

Back-End

Express.js

GraphQL

Mongoose

MongoDB

Ideation

Subtle was my first full stack project I worked on. While being in BrainStation, I had the privilege to learn in an environment surrounded by multiple “tech” disciplines. After networking with different classes one subject that caught my eye was Data Science. I was always interested in AI and machine learning, but I’ve never known the process or had a higher level of understanding on how it worked. I decided to pursue my capstone project with this in mind. Although I didn’t have the capabilities or resources to train my own model, I wanted to see how I can map out a model of data on my own. We were given two weeks to complete this project

Aside from the tech, I wanted to showcase a final project that reflected my favourite interest, dancing. There were only two weeks that I had to design, research, and code my project so I had to make quick decisions on what I actually wanted to make. gh 3/4 the course, I broke down each aspect of dance.

At first I wanted to have an application that broke down and analyzed one’s dancing. However, working with detailed motion and video data points were out of my current capabilities. This is why I broke it down to 3 levels. Dance, Pose, and Pictures.

Tackling the issue one by one helped me get started and initiated my project by focusing on pictures.

Key Challenges

BrainStation always pushed independent learning when it came to Web Development, and I realized it's one of the most important assets to have.

For my project. I had to venture and find the perfect libraries I could use for Subtle and overcome these challenges.

1. GraphQL | Being used to SQL and computational programming, I wanted to use GraphQL as part of my backend. As a GBDA student, it’s hard to find opportunities to code and there’s no opportunities to learn about Object Oriented Programming. GraphQL taught me a lot about data organization, and a new definition to perceive Objects

2. P5.js | Having learned creative coding in GBDA, I was excited to find a processing version for javascript. I was already familiar with this library, and they provided a great play- ground to test out. However, I did struggle quite a bit integrating media frames into my project.

3. Files | I had no idea how to code a file upload interaction. After getting the pop up to appear which apparently was the easier part, it was a mess of asychronous code I had to manage for it to function within my applications. Further more, I realized I was un- aware how phot data is used in the code.

Demo Day

After two weeks, demo day arrived and I was extremely nervous to showcase my project. BrainStation had invited industry recruiters and BrainStation alumni to help us network with professionals.

The reaction my project got from the staff and the guests were overwhelmingly positive. I had the opportunity to connect with employers from Vega, PeaceGeeks, Amazon, EA and more! The BrainStation experience gave me the confidence to know that my career was just beginning.