Graphical Representation of Sentiment
Dashboard (Displaying Leaderboards)
Diagram of Workflow
My inspiration comes from my first person perspective of how much food waste is occurring, even though it can be stopped with initiative.
What it does
Uses Clarifai's Food Vision API to find all the foods in every trash can picture created by my IoT camera. IoT camera takes one picture every 10 minutes, and will be turned on during lunch time. After lunch time is over, the device will send all the photos to my cloud app. Using all the foods found, my cloud app will display a leaderboard of which foods are being wasted the most in schools. Through individual school forums, I will compact all the ideas and form an overall sentiment pertaining to each school, using Microsoft's Cognitive Services (Sentiment Analysis). Two leaderboards, Best Sentiment Schools, and Worst Sentiment Schools, will also be displayed for users to see. Users can also see a graphical representation of a certain school's sentiment as months go by. Upon receiving these pieces of data, users can complete polls of which foods they would like to see in their menu.
How I built it
Used c9.io for a test environment to run my application, then uploaded to Github, then deployed onto Azure. Refer to the "It's Built with" section for a list of technologies I have used.
Challenges I ran into
First challenge was to learn how to write programs on Raspberry Pi, since I haven't done that before. I knew Python, so it wasn't a huge learning curve. Second, I was unable to connect to the WIFI at Hacker Dojo, since it used a portal system for authentication and due to the fact that I did not have a monitor/HDMI cable to connect to Raspberry PI. Nevertheless, I created a Python script which would control the IoT device and send data to my app. One major problem, was sample data, as I had to find many school forums already in existence, and images of food waste in trash cans.
Accomplishments that I'm proud of
I wouldn't be proud of the small milestones passed during the development of the app, more so the completion of development as a whole.
What I learned
I learned how to deploy a Node.js app onto Azure, and how to run scripts on a Raspberry PI.
What's next for Food Raccoon
Given more time, prototyping the shape/form of IoT device would be nice to do. One more functionality would be nice to be added; a list of all the new menu changes for recommending other schools what foods are better to add.