Goldman Sachs Portfolio Risk Analysis

Inspiration

A majority of our team were first time hackers, and we wanted to tackle a project that was challenging, put what we learned in ENGR 102 to the test, and helped us to learn new skills. We knew that we wanted to do a company challenge and decided that Goldman Sachs would just be a fun challenge and a good start into hacking.

What it does

The program takes in the stock symbol and the amount of shares that the user owns of that stock from their portfolio and calculates different risk values. It outputs the stock price at that moment, the risk per individual stock on a scale of 1 to 10, the risk of the portfolio on scale of 1 to 10, the state of the market on a scale of 1 to 10, and the overall risk of the portfolio by combining portfolio risk and market health.

How we built it

We have both backend and frontend components in this project. The backend is done in Python, and handles the calculations needed for the risk analysis. We can break it down into different parts. The first part of the backend is the use of the Finnup API. From the Finnup API, we called multiple different values including stock price and the individual stock risk. From these values, we created an equation that calculates the value of the portfolio health without outside factors. The second part of the backend is a web scraper to determine the current health of the market. We chose to use a web scraper over just values of today's market because we wanted the project to be usable for the duration of the domain timeline and wanted it to pull recent data. From the scraper, we determine a value for the current market health. Finally, for the third part, we combine the values of the portfolio risk and the market health using a certain ratio to determine the final risk of the portfolio's investments and print out all the values we want the user to see. For the frontend of the project, we used HTML, CSS, and javascript to build a user interface. We used PyScript to interface the javascript in the frontend with the python in the backend. We have a text box that takes in the stock symbol and the number of shares and calculates it. We are going to host the project using Domain.com so the project will be an actual risk analyzer that can be used by the public.

Challenges we ran into

We ran into many different challenges along the way. At first, we were doubting whether we even had the technical expertise to implement what we were doing. We had little understanding of all the financial concepts that we had to look at and didn't know how to implement many different things. We decided to split up the work and just pushed until we finished. One of the biggest challenges was interfacing between python and javascript. After 6 hours of working, one of our team members was able to figure it out. We also had many challenges with the final integration. We were in integration hell at one point trying to combine the backend with the frontend, but pushed through and was able to complete it.

Accomplishments that we're proud of

We are very proud of completing the project. At first, we were regretting picking the Goldman Sachs challenge because there would be a ton of information that would need to be collected and we needed to learn and pickup many new skills. However, with pure grit and determination, we were able to overcome these doubts and challenges and were able to create the project the way we intended it. We are also proud of the skills that we have learned in the last 24 hours.

What we learned

We learned a lot from competing in this hackathon. Many of our members learned new programing skills, such as javascript, web-scraping with python, creating websites with HTML and CSS, and interfacing between javascript and python. We also learned a lot of financial concepts such as expected return rate, the risk factors in a portfolio, and gained a better understanding of the stock market and stock portfolios.

What's next for Goldman Sachs Portfolio Risk Analysis

The next step for the project would be to add more risk factors that make the model a little more accurate. One idea that our team threw around but had no idea how to implement was creating an ML model that would look at the last 10 years of data from thee market and predict future data. We first started seeing this project as something for or first hackathon, but now we realize the real world use that the project we created could have.

Built With

• beautiful-soup
• css3
• finnhub-api
• html5
• javascript
• pyscript
• python