This iOS app predicts Fortune 100 companies stock price evolution through sentiment analysis. It uses SwiftUI for the layout, CoreML for the models creation, Twitter and News-api as the sources for data analysis.
The latest versions (1.4 and up) use Core Data to let you add, manage and save your own companies as well as cryptocurrencies.
For this app, I am using Twitter and News-api as the sources for my sentiment analysis.
I am performing the company analysis via 120 twitter comments about the company. Those comments are splitted in 2 sets. The first set gathers the 60 most recent tweets about the company (Ex: @apple) and the second set gathers the 60 most recent tweets about the stock (Ex: #AAPL). As the wording between those 2 sets of comments is very different, I used 2 different models trained on different datasets: IMBD dataset of 50k movie reviews for the first model and the Kaggle Sentiment Analysis on Financial Tweets dataset for the 2nd model.
Then, I fetch the 20 most recent news articles about the company and use the first model to perform the sentiment analysis on those news articles
Finally, I calculate a total score based on the sentiment analysis of those 140 comments
Swifter package from https://github.com/mattdonnelly/Swifter
Icons made by Icongeek26 from www.flaticon.com
I used the Circle Control code from this project: https://medium.com/swlh/replicating-the-apple-card-application-using-swiftui-f472f3947683
You can check the App Landing Page for a complete presentation and the Privacy Policy Page for further details on how your data is handled by the application
I provide the entire source code of this dema app for free. This demo app is licensed under MIT so you can use my code in your app, if you choose.
However, please do not ship this app under your own account. Paid or free.