NeuroFi

• 

  NeuroFi

NOTE: The Muse S EEG Headset provides nothing else besides raw electrodes data from the AF7 AF8 TP10 TP9 Sensors

Inspiration

We were brainstorming FinTech ideas when we realized how emotions, especially stress, impact financial decisions. Our dads' experiences with investing while stressed highlighted the need for a solution that helps people make rational choices rather than emotional, impulsive ones. When researching, we found that stress is a large factor in decision making and if you are too stressed, it could lead to irrational decisions. Creating a financial platform that can identify stress and help make suggestions on stocks felt like the optimal idea due to its potential large impact. With our platform, people would be able to make smart decisions when in the right mind and have an AI chatbot push towards less risky equities when stressed.

What it does

NeuroFi is an easy to navigate, prioritizing user experience, FinTech app. NeuroFi tracks stress levels in real-time and integrates them into financial decision-making. It recommends trades based on emotional stability, provides a news analyzer, offers financial advisor chatbot support, and visualizes stress trends through an emotional state chart and annual tracking, just using a commercial EEG. The trade recommendation comes from the financial advisor chatbot LLM trained on if stress is high, recommending less risky entities and when stress is low, recommending equities with higher risk.

How we built it

We developed NeuroFi as a web-based finance platform that uses EEG data, AI-driven market analysis, and real-time financial insights to help users make more informed decisions. We used a CNN to predict the users emotions and XGBoost to determine stress and boredom score. Along with that we implemented an Artificial Neural Networks and Random Forest Classifier to analyze news articles getting the sentiment, bias and the key points. We also developed a chatbot using the Gemini LLM and some prompt engineering. The EEG data was also derived from the PYLSL Library into five different emotions besides stress and boredom levels just off of the brain waves.

Challenges we ran into

1. Integrating EEG data with financial analytics.

2. Creating a seamless user experience while balancing data-driven recommendations with emotional tracking.

Accomplishments that we're proud of

1. Successfully linking emotional states to financial decision-making.

2. Building a functional prototype that tracks stress and provides investment insights which uses a LLM, using the data from the brain waves.

3. We trained all the neural networks on our own. We did the data collection, date preprocessing, model creation, and model training all by ourselves.

What we learned

1. The significant role stress and emotions play in financial decisions.

2. How to integrate neuroscience with finance to create smarter investment strategies.

3. The importance of user-friendly design and accessibility in FinTech solutions.

What's next for NeuroFi

1. Expanding compatibility with wearable EEG devices for real-time stress tracking.

2. Enhancing the AI financial advisor for more personalized investment recommendations.

3. Partnering with financial institutions to democratize smart investing for everyone.

Github Link Below! Github Project Link

Built With

• ann
• cnn
• csv
• flask
• javascript
• llm
• machine-learning
• numpy
• panda
• pylsl
• python
• random-forest-classifier
• react
• rnn
• typescript
• xgboost