Drive Volatility Indicator (DVI)

Using LLM and transformers to determine the social volatility of stocks

Comment

💡 Inspiration

Social stock volatility arises from the increasing impact of social factors on the stock market, a trend amplified by the rise of social media and the growing influence of public opinion. This phenomenon is driven by platforms like Twitter and Reddit, where information and opinions about stocks can spread rapidly, influencing investor behavior. Additionally, the surge in retail investing, facilitated by online platforms, means individual investors, often swayed by social sentiment, can collectively drive significant stock price movements. Emotional responses to news, rumors, and hype, accelerated by the modern, fast-paced news cycle, contribute to this volatility. Furthermore, public figures and influencers can greatly impact market sentiment, leading to quick and substantial shifts in stock prices.

💻 What it does

Our web application allows users to view 2 different stock volatility predictions. One built on top of standard economic indicators and the other on is predicted using sentiment analysis on how news articles and social media is talking about the company.

⚙️ How we built it

LLM transformers
Application Purpose
Streamlit Front-end web application
Programming Languages Python, CSS, HTML
Neural Network Machine Learning
Figma, undraw.co Design

🧠 Challenges we ran into

  • Collecting data from a variety of sources
  • The creative process to develop a solution to a problem that is some what achievable with our current skill set and time frame

🏅 Accomplishments that we're proud of

  • Implementing a working and functioning prototype of our idea
  • Designing and developing a minimalist and clean user interface through a new UI library and reusable components with a integrated design

📖 What we learned

  • how to use different libraries and API's like Beautiful Soup, Insomnia, and CSS
  • How to work together and collaborate with developers
  • How to use Streamlit to develop a fully-featured web application that users can access and interact with

🚀 What's next for DVI

  • Improving the data collection and storage process
  • Further developing the features and models used to make the predictions
  • Would like to make the process of predicting public so that users know where the numbers come from

Built With

Share this project:

Updates