Inspiration
Mental health is the wealth of the individual, but of recent, the COVID-19 pandemic and the resulting economic recession have negatively affected many people’s mental health and created new barriers for people already suffering from mental illness and substance use disorders. The best way to hear and understand the plea of the people as a whole within the digital age, is through social listening and natural language processing.
What it does
This module aimed at BEGINNERS, scrapes twitter for the most recent 400 tweets relating to mental health issues, it is built on existing python and Pytorch modules to analyze tweets and generate a word document report of what topics that relate to mental health are being mentioned, which places, people or things are being mentioned in relation to mental health issues and a dataset labeled by topic that can be used by mental health professionals, inclusive of sentiment per tweet to give the reader an indication of the emotion behind the tweet. Reporting for mental health related topics from social media (twitter) is made easy. Just 2 lines of code returns a dataset and report on the mental health related topic of choice. Import mmtweet mmtweet.mindmatters("USER ENTERS TOPIC RELATING TO MENTAL HEALTH")
How we built it
Using our knowledge of python
Challenges we ran into
Struggling to find a good library for tweet scraping using python.
Accomplishments that we're proud of
A working demo, although just executable scripts.
What we learned
Discovered flair (Pytorch based NLP module) for named entity recognition.
What's next for #MINDMATTERS
A mental health issue classification model and a proper interface.
Built With
- flair
- lda
- machine-learning
- matplotlib
- natural-language-processing
- python
- pytorch
- sklearn
- vadersentiment
Log in or sign up for Devpost to join the conversation.