What if your data sang to you? What would a Tinder for news look like? What if you had a recommendation engine powered by your past self?
Those are some of the ideas that surfaced at the third installment of hackingjournalism.com, an event held last weekend that brought together developers, designers and journalists and was organized by Embedly and MIT’s Future of News initiative, among other sponsors. Held at the offices of The Washington Post, this installment of the hackathon looked at data science: what it means and how it can help journalism. Storybench‘s Aleszu Bajak attended and compiled a list of the projects. For more coverage, see Matt Carroll’s blog and search Twitter for #datscihack.
What if a newsroom could see what was trending on the Internet and automatically surface archived stories related to that chatter? Dig.It mines Twitter Trends for keywords and then searches archived articles matching those keywords. Github here.
Great idea: Automating the surfacing of archived articles.
Who better to recommend you a story than yourself? oldNews is a recommendation engine that is powered by metadata from articles you’ve stored in applications like Instapaper, Pulse and Pocket to serve you stories similar to what you’ve read in the past. Github here.
Great idea: Serve a reader updates to a story she was interested in years ago.
If memes and clickbait are the fried chicken of news—the stuff you know is bad for you but click anyway—how do you get people to eat their broccoli? Renaissance Broccoli is a dashboard that recommends stories that may have lots of social shares but relatively few page views (this phenomenon being a measure of issues that are salient but not getting enough reads). Github here.
Great idea: Resurfacing high-quality stories that are getting low attention.
What if data could be converted into sound instead of a chart? Audiolyzer mapped historic data from shark attacks to tones with the frequency of the tone rising with higher datapoints. Github here.
Great idea: Adding an audio component to data visualization.
If a reader is sent a news article, how can they find additional news articles that supply additional, new information? Full_Story is an application that identifies keywords in the original article, looks on Twitter for similar stories on those keywords and then extracts paragraphs from those new stories that are the most dissimilar from the original article. Github here.
Great idea: Improving a reader’s understanding of a topic by automating the process of finding the most diverse coverage of an issue.
Content Magical System
If there’s one thing digital reporters can agree on it’s that Content Management Systems suck. Content Magical System is an email-based CMS that facilitates communication between editors and reporters and integrates with the WordPress API. Edits are tracked in a dashboard. Github here.
Great idea: Trying to eliminate the frustration reporters have with CMS’s.
Aren’t news websites a little too cluttered with social share buttons and newsletter signup forms? DSCN optimizes the type and number of call to action buttons shown with an article by relying on a user’s past browsing history, whether they have a subscription, and which social media platform an article is performing well on. Demo here.
Great idea: Eliminating Pinterest buttons for readers that aren’t on Pinterest.
Like Medium.com’s highlighting and annotation tools, Quotable wants to store and share your highlighting, sharing and commenting activity. Keywords are extracted from those highlighted excerpts to recommend additional articles. Github here.
Great idea: Saving highlighted passages in a database.
Reporters could use a platform to help store and organize articles around a topic they’re researching. FileThis does that and extracts names, places, tags and summaries to help search through content and filter what you want and don’t want. Github here.
Great idea: A platform to help organize research that goes into a story.
Want a reading application like Flipboard that learns what you like and don’t like and knows how much time you have? Flinder/Tinboard is a mashup of Flipboard and Tinder that learns your reading preferences as you swipe left and right and can surface content based on the amount of time you say you have. Github here.
Great idea: Tinder for news.
How can journalists find new sources and better understand the stories their readers want? The Bridge is a dashboard that compiles popular hashtags and retweets from Twitter handles that you enter. Github here.
Great idea: Diving deep into conversations on Twitter to find new sources.
Make Me Smart
Can press secretaries in Congress be helped to understand the context and research surrounding an issue? Make Me Smart is a dashboard bringing in state and national news, tweets, geographic trends and government reports on a given issue. Github here.
Great idea: An automated dashboard to understand the conversation surrounding an issue.
What if journalists could know every other angle that’s been taken on a story they’re writing? Journofriend aggregates coverage on an issue from other sources to find more diverse angles on the story. Github here.
Great idea: Helping journalists understand all the different angles other outlets are taking with a story.