What was the inspiration to create the News Bank?
There are tons and tons and tons of news articles out there. This amounts to an unimaginable amount of data that, the News Bank believe, could be used to predict stock prices and company behaviors (a.k.a predict the future). NewsBank would like to contribute and improve stock predictions using machine learning, web scrapping, and sentiment analysis.
What does News Bank Do?
NewsBank searches for a wide range of news articles from a myriad of trustworthy websites and different reliable news sources APIs on the internet to get enough data to confidently predict the outcome of stock prices and company behaviors. News Bank also shows a positivity/negativity score related to each news article, which contributes to and improves the overall algorithm behind News Bank.
How was News Bank built?
The News Bank was built using a Flask server written in Python in the back-end. It uses a news article API together with web-scraping techniques to gather and process a huge amount of news articles. The bank uses Google Cloud's NLP sentiment analysis and AutoML Tables to confidently and beautifully assess the positive and negativity scores of each scrapped news article (which, as explained above, will contribute to predicting stock prices).
Challenges the News Bank came across:
It is very challenging to gather such a huge amount of data, especially considering the small time frame the News Bank team has to work and scrape the web. Also, News Bank works on top of a myriad of technologies that need careful attention and proper management. So, the team has been attentively plugging and unplugging cables on the back-end and, at the same time, trying to provide the best possible user experience on the front-end.
Accomplishments the News Bank is very proud:
The News Bank came to life in such a small amount of time (Who would've thought this would be possible?). That is an accomplishment in itself. Also, as stated above, it was very challenging to gather a huge number of news articles from different reliable sources during the time frame we had to work. So, managing to create this huge dataset for our stock price and company behavior predictions is definitely an achievement that the News Bank team is very proud of.
What did the NewsBank learn?
The News Bank team learned better web scraping techniques so that the News Bank team could better search the web for more reliable news content. The News Bank team also improved its front-end knowledge by using news display and grid systems.
What's next for News Bank
News Bank is looking forward to improving its reach within the business and finance realm. We are looking forward to not only predicting stock prices, but specific company behaviors (like new product release dates, product features, and the list goes and on). The News Bank team is also looking forward to expanding its dataset of news articles in order to constantly improve the bank's algorithm.