CyberStock

CyberStock is a stock application that was built for my Masters project. This application has the primary features of a stock application similar to Apple’s default “Stocks” app. One of the key difference of CyberStock and other stock applications is that CyberStock incorporates machine learning. I built a RNN model in python utilizing Keras and Tensorflow. This model would read in large .CSV files of historical stock data and then output a prediction of a stock’s closing price. After testing, I found this model performed well with predicting stocks two weeks into the future. With this I converted my model to a CoreML model utilizing the coremltools package in python. I was able to incorporated my model into my iOS application using the CoreML framework, which allows a user to select a stock and see the two week prediction.

This app was built using Swift and UIKit with the MVVM design pattern. The application has different controllers that used standard UIViews as well as UITableViews. In order to feed in stock data, I connected the application to Finhub’s API. This API not only allowed for realtime and historical stock data, but also gave access to market news about each stock.

If you would like to see how this works, check out the application on the app store here! If you would like to see the source code, check out the repository here!