Anything you want. We’ve listed some Dream Projects posted by regional media groups for you to hack. To form teams and exchange ideas about these, go to our attendees page (sent in subsequent email). Email michael[at]majorplanetstudios.org with any questions.
CENTER FOR INVESTIGATIVE REPORTING PROJECT (I/III)
Type of request: Data analysis
Idea: Political consulting firms often create scores of committees to support or oppose a cause, leaving little discernible evidence about who is behind them — except for their mailing addresses. We want to use federal campaign finance data to help people uncover the hidden backers of these committees. By finding the addresses with the largest number of associated committees, we can determine the consulting firms behind them and create a resource for journalists trying to uncover astroturfing campaigns.
Application: Find stories and create a resource to help reporters
Contact: cdavis@cironline.org
CENTER FOR INVESTIGATIVE REPORTING PROJECT (II/III)
Type of request: Information extraction
Idea: A simple system to extract direct (and ideally indirect) quotations from news articles, along with their attributions, can have some very cool applications. For example, we created a project in 2010 called Politics Verbatim, which catalogued promises, predictions, attacks, statements of fact and other quotes from California’s two major-party gubernatorial candidates and turned them into a resource for voters. Problem was, we did it manually. Automating the process would enable this project and others like it to scale a new level.
Application: Build a tool to help hold politicians accountable to their rhetoric.
Links: http://www.politicsverbatim.org
Contact: cdavis@cironline.org
CENTER FOR INVESTIGATIVE REPORTING PROJECT (III/III)
Type of request: Web application development
Idea: Open records requests are the bread and butter of investigative journalism, but with all of the other demands on reporters’ time, keeping track of them can be difficult. A tool that would help reporters keep track of their requests — cataloging them, sending reminders when statutory deadlines pass, logging government agency response times, etc. — would be extremely valuable to the investigative reporting community. A social component that would allow reporters to collaborate would take that idea even further.
Application: Build a tool to help reporters track open records requests.
Links: http://www.politicsverbatim.org
Contact: cdavis@cironline.org
NEWS BAR
Type of request: Create a real, not cyber space, for consuming and sharing information
Idea: Theres a worm in the rise of the reinvented world. “Social media” is not very social. The lonely obsession with screens and interactive by myself applications can only do so much. There larger and much more challenging task is the creation of an engaged audience (this ought not be a forbidden word. Jay Rosen is profoundly wrong here) an engaged public conversation that is real and enjoyable. The News Bar is a fully out of the box idea that proposes a sustainable actual space where we reclaim the simple joy of reading, watching and listening to news and then we actually get to talk to real human beings about it. The business model: sports bar, but with news. Yes we sell alcohol and food and hire out of work journalists to mix drinks, integrate video and web to create dazzling environment for people who like news and the people who like people who like news. I have a proposal.
Application: Free standing, self supported business operated for profit.
Links: None
Contact: Calo@berkeley.edu
FACEBOOK SCRAPER
Type of request: Scraper to create timelines from social media feeds
Idea: I’d love a scraper that could page through a Facebook account to which a source has granted access and create a timeline of entries. Could wind up as simple as data dumped into .csv file or something more graphic and ambitious. Same with twitter, etc.
Application: “Would help reporters create chronologies for stories and make connections between events in people’s lives and real world activity. Could also be used as data viz in telling long, complex narrative stories — a personalized Storify. The tool could be expanded to also create social network analysis between Facebook connections versus Twitter connections.”
Links: None
Contact: t.christian.miller@propublica.org
ARTICLE COMMENT TOOL
Type of request: Web Service
Idea: Design and build a tool to enable commenting on/in a article. This would facilitate conversations about article content where it is relavant and contextual. A bonus feature would be categorization of the comments.
Application: Commenting tool for articles.
Links: None
Contact: KQED
BP-KATRINA PAYOUTS
Type of request: Application development; data search, analysis, extraction.
Idea: Use the BP oil spill class-action settlement process, to expose how some business-ownerships that qualify for cash settlements, have been systematically stolen from minors and incompetents. Show how same corrupted systems have been used since Katrina. And show how governmental and advocacy structures, nationally and on the Gulf Coast, and connections between Katrina and BP oil spill, need to be addressed.
Links: http://www.bit.ly/q5Cunx
Contact: Robert Evans
PULSE IMAGE FINDER
Type of request: Image Finder
Idea: Pulse make news reading much more fun and enjoyable by presenting content from publishers in a grid of stories. The essence of what makes Pulse pleasing is images inside articles. Articles with image look awesome, draw your attention; articles without images look a little boring and subdued. I wish someone could create an engine that takes in an article and spits out an image that might be appropriate to represent the article. Images could be pulled in from anywhere on the web, flickr, stock photos etc.
Application: Most of the content user’s see inside Pulse is served from our servers. If we had such an awesome engine, we would instantly add images to sans-image stories. This would enhance millions of users’ news reading experience almost overnight.
Links: None
Contact: ankit@pulse.me
RETRIEVING CONTEXTUAL INFORMATION
Type of request: Data Request
Idea: Pulls relevant contextual data regarding a particular story. For example a story about the violence in Syria could bring up information about Syria, a map, etc
Application: It could be used while reading an article
Links: None
Contact: albert@pulse.me
OUTSIDE SPENDING SCORECARD
Type of request: Application
Idea: An app that pulls data on super-PAC spending for/against candidates and mashes it up with the results of their races to provide a glimpse of whether super-PAC spending correlates with those candidates winning or losing. (We are currently doing this by hand.) The super-PAC data is all available via API (Sunlight Labs). Not sure about the election result data (from primaries so far).
Application: Find stories and create a resource to help reporters
Links: None
Contact: dgilson@motherjones.com
2012 MONEY TICKER
Type of request: Widget
Idea: A widget you could put on a news site’s front page or embed in articles that provides the latest numbers on:
• Presidential fundraising: Total, Obama vs. Mitt advantage
• Congressional fundraising: Total, Dem vs. GOP advantage
• Super-PAC fundraising: Total, liberal vs. conservative advantage
Application: Display total fundraising to date in real-time
Contact: traja@motherjones.com
POP-UP VIDEO STYLE CAMPAIGN ADS
Type of request: Playing with popcorn.js
Idea: Crowdsourced popup annotated campaign ads and speeches using HTML5 video and popcorn.js — layer popup-video-style funding facts (with links and sourcing) into video of candidates talking; take and vet user submissions.
Contact: traja@motherjones.com
HEALTH CARE EXCHANGES NEWSGAME
Type of request: Data Request/Newsgame
Idea: We are going to be launching a project to look at state health care exchanges. The questions we will want to ask are: who will be covered? who will benefit (i.e. will it be set up to benefit humans or prescription companies, etc)? how will the exchange be funded? which exchanges offer which services?
Application: We would like to crunch the data and turn it into a news game of sorts (i.e. if I know I need to have a heart bypass, am I better off living in Connecticut or Alabama)? I know we will have to crunch a ton of data, but I’m very new to all this and don’t know exactly what to ask the programmers to help us do. Only a few of these exchanges have been set up as of now (Massachusetts, California). The dilemma is that they are state run, so each will have a different way of organizing the data.
Contact: joellen@themediaconsortium.com
Title: BANTR (Bay Area News Tracker)
Type of request: Mobile app
Idea: We propose to gather original reporting from dozens of disparate nonprofit newsrooms from around the Bay Area into an indispensable local public media mobile app that consumers will actually pay for. BANTR will aggregate and promote breaking and enterprise news from nonprofit publishers in a userfriendly consumer-facing interface to compete for “best local news app.” The San Francisco Public Press has forged content-sharing and reporting partnerships with more than 30 local public media and civic organizations. BANTR will extend this network to a platform that offers a new revenue stream for all partners.
BANTR helps solve the problem of fractured branding among local public media. It allows nonprofits in the Bay Area to turn extensive existing collaboration networks to the advantage of the sector in a competitive news environment. A clean, lightning-fast user experience with dozens of fresh local news stories daily will give the dominant, commercial news sources a run for their money. The platform can be adapted for other public media markets across the country. We will offer it to the 60 members of the Investigative News Network, which has already standardized members’ news feeds for non-exclusive syndication.
Application: We plan to charge modest subscription fees for the smartphone and tablet versions of the BANTR app. One-third of revenue will go toward product improvements, one-third will go to pay editors, and the final third will be divided equally and distributed to participating partner organizations.
Links: None
Contact: Lila LaHood — lila@sfpublicpress.org
VISUAL SCRAPER
Type of request: Data analysis/information extraction
Idea: With the demise of Google’s Needlebase Web scraper, data journalists looking to capture online data from complex government databases were left with off-the-shelf products that are either inferior or outrageously expensive. While learning to code individual scrapers by hand is an option made easier by tools like ScraperWiki, there is a serious need for an intuitive, visually based Web scraper for reporters lacking the technical know-how. An inexpensive, user-friendly tool would greatly expand the data-based reporting capability of small newsrooms and citizen journalists alike.
Application: Design and build a Web scraper that allows reporters to capture online data via an intuitive user interface
Links: Background here: http://www.reporterslab.org/needlebase-and-the-future-of-scrapers/ and here: http://www.reporterslab.org/needlebase-and-the-future-of-scrapers/. Test data here: http://reviews.reporterslab.org/search?q=&type=products&category=scraper-2011-09-28
Contact: Tyler Dukes, tyler@reporterslab.org
INTERESTING DOCUMENT FILTER
Type of request: Document management/content analysis
Idea: Reporters who regularly submit Freedom of Information Act requests are used to drowning in documents, whether it’s banker’s boxes full of printed pages or massive PDF files containing hundreds of individual items. Taking the time to sort through this content is a major drain on a news organization’s resources. But by using a machine learning-powered application with an intuitive user interface to highlight the most newsworthy items based on the user’s interests, a reporter could guide his or her exploration to the most promising material in a massive document dump.
Application: Design a machine learning algorithm that uses training data like emails and story archives to prioritize items in a document dump, helping reporters maximize their time
Links: A Stanford project called Muse can already use training data from your emails to highlight interesting things on web sites: http://mobisocial.stanford.edu/muse/. Potential test data here: http://reviews.reporterslab.org/ and here: http://www.muckrock.com/
Contact: Tyler Dukes, tyler@reporterslab.org
