Saw paper about using keystrokes to detect alzheimers. Right now Covid-19 is making more people work from home, and I think negative sentiment will be rising.

What it does

Listens in background for sentence termination. Gets positive & negative sentiment from API. Then it saves daily high and low for 30 day running average for positive & negative sentiment. Then these sentiment data are plotted on a graph to see trends.

How we built it

Started with python to watch for keystrokes in background, got the sentiment scores with expertai and then saved to a db and plotted with Matplot .

Challenges we ran into

Was new to keyboard library so learned it had function to convert key sequences to words and lacks the ability to listen for multiple sentence termination keys, .,?,!.

Accomplishments that we're proud of

Every day at midnight calculating moving 30 day sentiments.

What we learned

Mostly about calculating moving averages and plotting them.

What's next for MoodBoard

Breaking daily average down to hourly, listening to all sentence terminations and sharing sentiment scores with others. With more typing data we can start to develop suggestions around mental health and better understanding dementia.

Built With

Share this project: