Inspiration The inspiration behind Mood Journal stems from the growing importance of mental health and emotional well-being. Often, individuals are unaware of their emotional patterns, which can prevent them from taking proactive steps to improve their mental health. Mood Journal was designed to provide an easy and automatic way to track moods based on conversations and provide insights that can foster self-awareness and growth.
What it does Mood Journal is an AI-powered app that processes real-time audio conversations, analyzes the sentiment, and categorizes emotional states. It stores mood data securely, allowing users to reflect on their emotional patterns over time. With a simple interface, users can track positive, negative, and neutral emotions, helping them understand their mental state better.
How we built it We built Mood Journal using:
Backend: Flask for handling webhook requests and routing. Sentiment Analysis: NLTK's VADER SentimentIntensityAnalyzer for emotional classification. Database: SQLite for securely storing user moods and transcripts. Testing: Postman for API testing and debugging. Hosting: Local Flask server (can be extended to external hosting with tools like Ngrok). The app was designed to process real-time JSON data from Omi devices, analyze sentiments, and store them in the database. A simple /moods endpoint allows users to retrieve and view their logged moods.
Challenges we ran into Sentiment Classification: Negative statements with positive words (e.g., "not good") were sometimes misclassified. This required fine-tuning our sentiment analysis logic. Webhook Testing: Setting up a local environment to mimic real-world API calls took additional effort and required tools like Ngrok for external testing. Data Storage: Ensuring that moods were stored and retrieved efficiently in the SQLite database without redundancy or errors.
Accomplishments that we're proud of Successfully implemented a real-time sentiment analysis pipeline. Designed a simple and intuitive way for users to track and review their emotional patterns. Created a functional webhook that integrates seamlessly with the Omi ecosystem. Overcame challenges with classification accuracy by refining sentiment thresholds and logic.
What we learned The importance of accurate sentiment analysis and the challenges of natural language processing. How to work with Flask and SQLite to build scalable and efficient backends. The value of testing and debugging API endpoints using tools like Postman and Ngrok. How real-time emotional insights can foster self-awareness and promote mental well-being.
What's next for Mood Journal AI Data Visualization: Add graphical representations of emotional patterns over time, such as mood charts or trend lines. Integration with External Tools: Enable exporting mood data to mental health apps or journaling platforms. Mobile App Extension: Build a mobile app that directly interfaces with the Omi device for an enhanced user experience. Machine Learning Models: Train custom sentiment analysis models to improve classification accuracy and handle nuanced emotional expressions. Cloud Deployment: Host the app on a scalable platform like AWS or GCP for broader accessibility.
Log in or sign up for Devpost to join the conversation.