Telefinancial Forecaster

Inspiration

Even before I applied for my course in Business Analytics, I always wanted to build something related to screen stocks.

What it does

What it does is to keep a portfolio of stocks you choose to keep and is able to perform quick forecasting of future prices so that you can quickly make a decision to own or pass over ownership of the stock. It also provides a summary info about the performance of the stock currently.

How I built it

The backbone is built as a telegram bot which can run the various functions. Information retrieval is performed by yfinance, an unofficial Yahoo Finance library, time series analysis is done by statsmodels, and graph plotting is done by matplotlib.

Challenges I ran into

Initially, I was going to use d3.js, but was unfamiliar with its operations and I feared for the lack of time for me to know it well enough to use it. Afterwards, yfinance was also having problems as the core library has some bugs which were not addressed by the creator, but someone has pushed a bugfix for it, so I applied that.

Accomplishments that I'm proud of

Proud of having made a Telegram bot all by myself.

What I learned

Learned how to apply bugfixes to the library directly

What's next for Telefinancial Forecaster

Possibly transition it to a full mobile application, and incorporate further machine learning, especially neural networks.

Built With

• matplotlib
• python
• scikit-learn
• statsmodels
• telegram
• yfinance