Sentient Data Analyzes
Collects and analyzes data to give personalized data and data trends!
Project Overview
What is this app?
With any topic or phrase the application is able to collect data from a given time period using the requests and BeautifulSoup Libraries. The data is then ran thru the VaderSentiment Sentiment analysis model to get a Fine-Grained score. Using this score The app is able to give document recommendations and topic data analysis that mocks the stock market of the topic without using any numerical data.
Why vaderSentiment?
vaderSentiment was chosen due to the speed of using a lexicon / rule-based sentient analysis vs a ML one. By doing this the accuracy for individual documents was scarified in exchange of speed but it is balanced out by averaging out the data and any outliers would be drowned by the rest of the data.
Real world applications
Schools
This type of application can be used by students of all grades when doing research for a paper on a certain topic. This tool can provide the most relevant articles for the desired topic by eliminating moist of the bias around many political topics.
Law / Political Discussion
On the opposite note from this tool can be used to find the most bias articles around a certain topic for debates for example by a lawyer making a specific case.
Stock Market / Research
Because vaderSentiment is specifically attuned to sentiments expressed in social media this is a great tool for researching how the masses feel towards a subject at any given point in time. This application can be used to map out stock market trends without using any numerical data but purely online publications or getting public feedback on topics.
Nvidia Stock compared to Sentient analysis
Sentient model is able to chart stock market trends from posts on the internet
Recomended articles for the search "Ukraine vs Russia"
The model was able to recommend different articles based on the moral(Negative, Neutral, and Positive) after analyzing 199 articles in just a few seconds! Some key phrases have been highlighted to show the differences in the documents focus as well as an overall topic moral has been calculated for this topic.
Microsoft Stock compared to Sentient analysis
The model's analysis matched Microsoft's stock trends almost perfectly throughout the day of Friday Aug. 19. This could be further developed for an automatically day trading/ swing trading model.
