We were inspired by how the media is delivered to the public in today's world. The internet has became so complex and the amount of information accessible is outrageous, however, there is a tendency of the media presented to the public being simplified, and the view of the mass pubic becomes narrow. We thought of an entertaining way to tackle this issue, which is the user to be able to find polarized articles of the news topic of your choice.
What it does
Our web app is able to scrape the web for the article concerning a topic of your choice. It will then present two articles to you. One being the most 'positive' it could find, and the other being the most 'negative'. The values are based on our algorithms calculated with the Google Cloud NLP machine learning sentiment analysis: the article's magnitude values and overall sentiment score. The system is precise and really fun to interact!!!!!
How I built it
This program was built using python and then Flask for the web app. The sentiment analysis was done by utilising the google cloud API. We used the google custom search API to scrape the internet.
Challenges I ran into
A large challenge that we faced was scraping the internet without being stopped as our behaviour was not in line with that of a human. Also using tools such as the google cloud ML and flask was our first time, and we took a lot of time aside for tutorials and how-to’s.
Accomplishments that I'm proud of
Today we had used natural language API's saw the first time and we even trained a data set for the very first time. The precision of the product is great, and it is joyful to play with/ explore different ideas and perspectives.
What I learned
Over the weekend we have learn a great deal about APIs, natural language learning and much more.
Google Cloud NLP
AutoML Sentiment Analysis
Google Custom Search API