Inspiration
As people spend more and more time on the web these days, their lifestyles shaped around the sites they frequent. We wanted to use sentiment analysis and opinion mining to find the tone of the websites, and with the data, create an overall impression of the user's positivity/negativity values as well as the world's search values.
What it does
WebTone takes an url of a website, scrapes through it to find all the text in the page, compare the words to a lexicon of 7k positive/negative words, and scores it on the ratio of pos versus neg diction. Then the App animates the scale of positivity/negativity to graphically show the tone of the website and prints out the words it sensed. The customized graph bar with on the top of the screen is the world's recent web search sentiment analysis, and when tapped the average positivity/negativity value of the user is shown with a circular graph in the middle.
How we built it
We built a web scraper on ios to parse the text of any website, used an opinion lexicon to compare and analyze, then recorded the ratio of positivity. Then we designed and animated the circular scale to show different levels of opinions. User defaults is used to record the average positivity/negativity value of the user's searches. The backend of every web search sentiment analysis is saved in Firebase with a key/value pair and called whenever a new value is added, updating the programmatically made bar graph on every users' screen.
Challenges we ran into
We had time constraints as we started the project rather late in the hackathon, about twenty four hours in. However, WebTone really called out to us, and we're happy we stuck to it.
Accomplishments that we're proud of
Designing all the images and icons and animations.
What we learned
No matter how sleep deprived- % != / .
What's next for WebTone
Create a safari extension to go along with the app & expand lexicon.
Log in or sign up for Devpost to join the conversation.