Hack My Mood

Who's ever been down in the dumps because it was a dreary day outside? Many people struggle with mood swings and fluctuations brought on by changes in the weather. Everything from everyday feelings to Seasonal Affective Disorder and depression can be affected by changes in our environment. What if people had more awareness and control over the patterns in their mood and looked out for warning signs in the future?

Using the power of social media, natural language processing, and the wealth of weather information on the web today, we can mash up a variety of data sources. Using Hack My Mood, users can track the sentiment of their tweets over time, and relate that to the weather in their area that day.

We believe that awareness is the first step to better mental health, and by educating people on the patterns that arise in their own writing, we can build a happier, healthier world.

APIs Used:

  • IBM-Watson Alchemy API

  • Weather Underground

  • Twitter

Results:

At least on my tweets, it does appear that sentiment drops slightly as temperature drops. A larger sample of data would be needed to verify this. I plan to run this analysis on a much larger dataset in the future.

Challenges:

Combining data from so many libraries caused a lot of latency, and there's not enough time to run through the whole process during the demo. Multiprocessing and caching would speed this up in the future.

Future Plans:

Using the Weather Forecast API, we can predict days when bad weather may be coming, and alert the user to be more mindful of his or her mental state.

Share this project:
×

Updates