News Article Sentiment Analysis
Project searches Google for news articles. Relevant passages are identified and used to summarize the person's profile. The collected articles are available on the dashboard hosted on Replit.
News Article Sentiment Analysis
Project searches Google for news articles. Relevant passages are identified and used to summarize the person's profile. The collected articles are available on the dashboard hosted on Replit.
The problem News Article Sentiment Analysis solves
An institution, organization, or person interested in gathering information about a person can use this product to analyze the sentiment expressed about that person in news articles, which is a very cost-effective way of gathering information. This product can also save manpower and time that would otherwise be spent on manually gathering this information.
Challenges we ran into
We faced two major challenges:
-
Scraping data within a reasonable time and cost: We switched from using Apify to using SerpApi + newspaper to make scraping more efficient. Although Apify is more accurate, SerpApi + newspaper was more efficient.
-
Extracting relevant passages from the news articles and obtaining an accurate model: To solve this challenge, we filtered the sentences containing relevant keywords and used the spacytextblob library for accurate sentiment analysis.
Discussion