With the recent political climate, we've all heard of the trending "Fake News," a type of yellow journalism that consists of deliberately spreading misinformation / hoaxes through both print + digital mediums. Fact checking became more important than ever, and along came many features and products to help consumers feel more confident, safe, and properly informed.
But the reports and articles covering that information have much room for "creative freedom."
So how is it affecting us mentally and emotionally?
More often than not, the rhetoric that is used to communicate these current events are often drenched in offensive, condescending, and targeted language; and we don't realize how much of an impact this has on a person's emotional / mental states and ultimately, long-term, everyday behavior.
Lead Product Designer
Tasks: UX Flow, Competitor Benchmarking
2 Days - CalHacks 4.0
How might we curate a feed that delivers news to the reader without rendering a negative, long-term effect on emotional / mental health?
A mobile application that aggregates and curates news articles based on an overall "Positivity Score."
- Are we trying to curate only "neutral" articles?
- Exactly how are we deciding the overall positivity score of an article / headline?
- People react and respond differently to certain articles & headlines. How do we make the feed personal and fit to the individual reader? What's the most efficient way to get their feedback on the "positivity score"?
- Isn't this essentially brainwashing? In that sense, how do we make sure we're not creating an echo chamber of articles, and making sure the important updates and current events still reach the reader?
- Does this curation start from filtering the news sources at which the articles are aggregated from? If so, how much autonomy should we give users to choose those types of sources? Will adding the news source add any implicit bias to the way the news source is read + internalized?
FTUE - Onboarding
First time users are prompted to sign-up using an existing Twitter / Facebook account to import authenticated data - geolocation, age, interests, likes, followers, and general activity behavior.
The user is then prompted to an interests screen in order to curate a personalized feed from the zero state landing page.
The categories will be based on the the most popular news categories from existing competitor apps:
• Apple News
• Associated Press
• NPR news
• BBC News
Home / Newsfeed
The home newsfeed is updated in real time, and prioritizes the "most positive" news story on top.
Each story is tagged with a positivity score alongside a color code system, and the feed would not only show the most positive news, but also take into account user's interests, variety of categories, and most pressing current events.
After user taps the story, they are led to the full article screen, and after scrolling to the bottom, users are prompted to input a positivity score on a spectrum of -100 to 100. This score will be added to the overall score given by readers, and this would update the article's total score on the feed.
Because the terminology might lead to confusion to a majority of users, we utilized the heart icon, and decided to prompt a "You gave __ likes" pop-up modal to show a direct correlation between liking a post and having a higher positive score. This idea was inspired by Medium's claps, replacing the ubiquitous like button feature and allowing a more qualitative "standing ovation / applause" experience. In hind sight, it might be more beneficial to include a full-page modal screen to explain the positivity score - so this may have a high education cost from the user's perspective.
How positive are you?
After reading a minimum number of articles, the application is able to generate a personalized positivity score - an average from all the articles read + scores given by the readers themselves.
By creating this score feature, we are adding a slight gamifying effect to encourage users to be more conscious of the articles they read, question why some articles are assigned that score, and be more conscious of the difference in rhetoric / content displayed in this app ecosystem vs. external platforms (secondary and competitor apps).
The score is also color coded into three distinct sections (relative):
• Red - not positive
• Orange - neutral
• Green - positive
And the ranges would be dynamic based on each reader's scoring behavior, and attaching those scores to the range of words / phrases associated with the article. This will be handled by the backend machine learning + TensorFlow libraries (beyond my scope / capabilities) - specifically with sentiment analysis and NLP.
Screens - work flow
Jerry Jin - Backend Database Developer
Austin Luong - Front End Developer, Android (Expo, React Native)
Laney Huang - Front End Developer, Android (Expo, React Native)