Here is how we want to combine AI with Stock analysis…

The Problem

Investing and portfolio management are key skills to grow one’s financial wealth and security. But today, beginner investors are intimidated by the complexity of investing. According to a 2018 study done by Gallup, “…the combined percentage of adults younger than age 35 with money in the stock market in 2017 and 2018 stands at 37%, down from 52% for people in that age range in [2006 and 2007]”.

What is FalconUp

Our FalconUp team is currently building an application to help new investors with limited experience trade with less risk, making trading more compelling for new investors. FalconUp is made with machine learning models that utilize previous stock prices and stock articles to account for price changes and general societal behaviors regarding the stock market. While numerical values are important, we also take into account how the stock market is heavily affected by the thoughts and emotions of society. We designed our model to be for the long run, as news and social media platforms will take time to affect stocks.

Trusting AI

In this day and age, people trust AI much more in terms of functionality, practicality, and predictions than humans. For instance, Chiara Longoni and Luca Cian from HBR conducted an experiment to see if people trust AI or sales peoples’ recommendation for a sample of hair care products. “As predicted, when passersby were asked to focus only on utilitarian and functional attributes such as practicality, objective performance, and chemical composition, more people chose the AI-recommended sample (67%) than the one recommended by a person.”

How we started

We started FalconUp without much experience in machine learning or how to make apps or even having met before becoming a team. Even so, we came up with an idea and began building it into an exceptional ongoing app over the course of the last year. We have also received valuable insights and feedback from generous individuals who gave us their time to improve our app in both functionality and aesthetics. We currently have four active team members who are all interested in coding and stock trading and an extremely helpful instructor who gives us a lot of assistance in building the app.

Progress

In the span of one year, we’ve progressed a lot. The portion of our system that users interact with is the React app, which we have built and is up and running. Our backend currently houses our data scrapers that gather stock prices and articles and our database which stores the information accumulated by the scrapers. These services run daily to maintain up-to-date information. The user interface is designed to be intuitive and easy to use. The design is constantly being updated to improve user experience. We were confident enough in our app to participate in an entrepreneur competition called the Diamond Challenge. Overall, our app came from absolutely nothing to something amazing that we can all see today.

Challenges

During the course of the project, there were some challenges that we knew we were going to face and some that we didn’t know about. For instance, the only people that knew each other were Jeffrey and Enzo, but even then it was still a bit hard to pass ideas around. One big challenge that we experienced was our team leader, Andrew, leaving. However, Jaymin Ding showed up, and he was the ideal teammate that we were looking for. Another challenge we faced was the fact that we were not familiar with how to code or how any part of the app was going to be made. Our instructor helped with that, giving us great ideas and showing us how we should do it.

Future plans

In the future, we plan to improve FalconUp in many ways, especially with the accuracy of our predictions. An idea we had was scraping data directly from social media such as Reddit and Twitter for big events, which we can use to create a better prediction model. Using articles and social media to represent that sentimental factor, our prediction will be considered to be more accurate than one solely based on numerical data. Additionally, we also thought of creating different kinds of machine learning models that would contribute to an increase in our accuracy. We also considered the idea of adding a payment system to start bringing in income to improve FalconUp in many more ways. In conclusion, FalconUp is still in progress but we have some ideas on how to make it better in the future.