DeltaVision

AI-powered investing platform designed with first time investors in mind. Relies on probability analysis to calculate the expected value of a particular financial decision.

Comment

Inspiration

DeltaVision was inspired by our personal struggles getting into trading. As beginners, we were faced with an overload of information and a scarcity of on-demand, high-quality and understandable analyses. Essentially, we lacked a centralized trading platform for inexperienced young traders that would address these issues. So, we decided to create our own platform that would help new traders make smart financial decisions regardless of their experience level.

What it does

DeltaVision has two functionalities. The first is to provide an accessible, beginner friendly platform to view information summaries about stocks and access digestible AI analyses. The second, packaged in a subprogram named EVCalc, is to calculate the mathematical expected value for any financial decision that a user might make. EVCalc parses user prompts, does research, generates potential consequences, and calculates and recalculates probabilities to come up with a final EV score for any financial decision.

How we built it

We designed the UI in Figma and then prototyped it for the web and mobile (HTML and CSS frameworks, JS).

Challenges we ran into

DeltaVision interface

  • Building out the design language so that the interface would be functional for both mobile and widescreen
  • Adding enough complexity for advanced users while keeping it simple for beginners

EVCalc

  • Designing a way to cut statistical noise for generated probabilities
  • Adapting the "Wisdom of the Crowd" statistical phenomenon for LLMs
  • Determining a list of "consequence types" to generate possible effects and probabilities for
  • Finding a mathematical formula, based in personal finance and economics, that would allow us to calculate expected value for a wide range of situations
  • Adding financial safety guardrails for the integrated AI agents
  • Standardizing a wide range of subjective consequences into one objective mathematical scale

Accomplishments that we're proud of

  • Coming up with a clean and easy-to-use UI for both web and mobile
  • Developing a comprehensive, AI-powered algorithm to calculate EVs and generate accessible reports for users
  • Using concepts and results from economics and game theory to build a consistent and powerful tool

What we learned

  • How to convert subjective decision prompts to numerical expressions to be solved
  • The importance of defining mathematical units and scales to solve subjective problems analytically
  • The balance between providing information to users and keeping interfaces simple and navigable
  • The importance of including reappraisals and checks in AI-powered algorithms to guarantee consistent results

What's next for DeltaVision

  • Actually implement ordering to real world exchanges
  • Getting clearance and insurance for investors so their money is secured
  • Adapting EVCalc to a financial advisor that can aid with personal finance decisions outside of stocks and bonds

Built With

Share this project:

Updates

Submission history