Inspiration
We were inspired to create this project because of the amount of propaganda and misleading information that can be found on the internet these days. Our project helps users identify the tone of whatever article they are reading so that they can be aware if the author of a certain piece has any kind of skewed view.
What it does
Our project scans an article and uses datasets of positive and negative sentiment words. It compares the words in the article to the dataset of sentiment words and counts the number of positive and negative words in order to calculate a sentiment score for the article. A google extension is then used to report the sentiment score and its meaning to the user.
How we built it
We started with the Stanford Treehacks template for word replacement in a chrome extension form. We knew we wanted to make some software to improve media literacy and understanding, so we created our own form of sentiment analysis to scan a webpage and produce a certain tone indicator.
Challenges we ran into
One challenge we ran into was working with ML sentiment analysis models. We tried a variety of premade and pretrained models, such as TextBlob, but with low levels of accuracy. In the end, we found it more beneficial to create our own set of terms to search for.
Accomplishments that we're proud of
We are proud of tackling a new type of development that both of us were unfamiliar with. Using chrome extension software required us to think about interactions between languages, data storage, and web page scraping in new ways.
What we learned
We learned about HTML/js messaging, sentiment analysis, and using chrome developer software.
What's next for Sentiment Report
We hope that MSR can be a tool to help the public consume media in an educated way. In the future, we could expand to include more data in the report, possibly involving stats on misinformation and fact checking.
Log in or sign up for Devpost to join the conversation.