Frontend for Shap Emotions Correction

This demo detects sentiment: positive (green), neutral (black) /negative (red)

the predictions update as you type

Explaining sentiment

You can also explain the predictions

Scores are presented on scale from blue to red. Red means that this word increased probability of this output class. Blue on the other hand decreased it. White words are neutral for this output.

How to run

1. Run sentiment API:

docker run -p 8000:8000 finloop/shap-emotions-correction-api:latest

1. Intall npm dependencies:

npm install

3. Run server:

npx parcel src/index.html