ChatCoach

💡 Inspiration

Truth be told, we're not the best presenters... Sometimes we cross our arms during the entire presentation, sometimes we start using lots of filler words. But we'd like to get better at it! Introducing: ChatCoach!

🚀 What it does

ChatCoach is simple. Click on 'Start' and present in front of your camera. After you're done, you'll be able to view stats and graphs based on all your skills. First, you'll get a ratio of the time you spent looking in all three directions: left, right and in the centre. You'll also know if you ever looked down at cue cards, or if you crossed your arms. Your audience wants your full attention and its through open body language and eye contact that you can achieve that. Finally, the application analyzes how often you use filler words so you don't have to. All the information is presented in separate graphs and then summarized on one page.

⚒️ How we built it

This web application operates using Vanilla Javascript. The video analytics were built using MediaPipe and the speech-to-text was integrated using the Web Speech API. We also used Chart.js for everything graph-related!

😢 Challenges we ran into

We had never worked with video intelligence and a lot of documentation (or lack of it) was read over the weekend. Moreover, our team hadn't worked without a front-end framework in a while. We had to go back to the basics on this one and, ultimately, we realized less is sometimes better.

🔥 Accomplishments that we're proud of

This project constitutes the first deliverable our team created that fully works! We're also super proud that we got to successfully work with MediaPipe. Quite frankly, we almost gave up when nothing seemed to work and the online documentation was nonexistent. But a few 'console.log's later, the movement tracking started appearing on our screens!

📖 What we learned

First off, we got to get off our React.js high horses and see how powerful simpler tools can be! We also got to play around with video intelligence and explored all the various use cases there are! Finally, we're now more knowledgable when it comes to good vs bad data visualization.

👀 What's next for ChatCoach

There are a lot of different avenues we can explore! Our project only showcases the surface of what our application can do. With a second camera, we'd be able to monitor posture and even alert users as they present to stand proud. We could also expand a lot more on our speech-to-text component, like sentiment analysis of the presentation.