hr/deepeval

Deep Evaluation of classroom sentiments (OxfordHack 2017)

DeepEval ⚡️

Development

Structure:

• master - production staging branch (with tagged releases)
• dev - main development branch
• {YOUR_NAME}/{ISSUE}-{FEATURE_NAME} - feature branch

You should work on your feature branch until you have implemented it at which point you land it onto dev via a merge or, preferably, a rebase. Once dev is release-ready it is landed onto master.

deepeval's Issues

Implement persistent connection between client and server

Use sockets to establish a persistent connection between the react client and the node.js server (via socket.io).

Capture and return analysed sentiments to client

Procedure
• Capture video stream
• Select ~1 frame per second from the video capture stream
• Send it off to the server-side
• Forward frames and get sentimental data from face API
• Process results (average motions)
• Store & send back the final results

Implement user interface

Pages:

• Dashboard
  • Overall emoji (satisfaction, since joined)
  • Overall emoji times series (since joined)
• Lecture
  • Overall emoji (satisfaction, of the lecture)
  • Emoji time series (of the lecture)
  • Order: fear, disgust, anger, contempt, sadness, neutral, happiness, surprise
• New lecture
  • Real time video feed
  • Real time emojis
  • Real time overall emoji