Kamis, 07 Maret 2013

Nieman Journalism Lab

Nieman Journalism Lab

Data science, commoditized backends, and the need to know code: A roundup of NICAR 2013

Posted: 06 Mar 2013 10:05 AM PST

nicar-2013-logo

The annual Computer Assisted Reporting conference, known colloquially as NICAR, wrapped Sunday. Of all the journalism conferences held throughout the year, this is the only one to specifically focuses on the needs and interests of reporters and editors who work in investigative news and data journalism. It’s a tremendous three and a half days of classes, panels and camaraderie. Attendees come to learn, share, and solve some of the most pressing issues currently facing the industry.

For the last three years, I’ve been collecting NICAR presentations, tutorials, tools, and work samples, because learning and looking through everything presented takes time. This year’s sessions were wide-ranging (including a brand new session on how to host high-traffic news apps), but there were still a few notable themes running through it all.

DIY

Many attendees are in the “people should learn to code” camp, and for good reason: With budgets gettings squeezed (at the start of one hands-on tutorial, an urgent question launched from the back: “Is it free?”), fee structures changing, and the growing desire to customize the look and interaction of published work, journalists’ need to literally and figuratively own what they make is more important than ever.

al-shaw-ruby-nicar

This year, there was much heavier emphasis on learning JavaScript, Python, R, and Ruby. Jeremy Bowers (NPR) and Serdar Tumgoren (Washington Post) posted their Fundamentals of Programming in Python materials to GitHub and created a Google Group for class members. Ron Campbell (The Orange County Register) and Christopher Schnaars (USA Today) offered Programming for the Rest of Us to those who wanted to code but were worried about the learning curve. (Though not specifically taught this year, there are also a number of excellent d3.js tutorials, which I’ve collected in the references section of my list.)

That said, lots of data still comes in Excel spreadsheets, and Krista Kjellman Schmidt (ProPublica), Linda Johnson (Lexington Herald-Leader), Denise Malan (Corpus Christi Caller-Times) and MaryJo Webster (St. Paul Pioneer Press) all gave terrific presentations on how to do it.

We’re sharing best practices

Sometimes it’s hard to tell when “best practices” are coming from experts or pretenders. In the case of NICAR, it’s a pretty safe bet that you’re getting advice from people who’ve tried, tested, and refined their methods.

xkcd-heatmap

Dave Cole (Mapbox), John Keefe (WNYC), and Matt Stiles (NPR) shared what what works well for mapping. Tasneem Raja (Mother Jones) and Sisi Wei (ProPublica) showed how to make interactives fun. Steve Myers (The Lens) explained workflows for social media discovery, verification, and publication, particularly during breaking news.

Learning from data science

drew-conway-data-science-venn-diagram

The oft-cited Venn diagram from data scientist Drew Conway describes data science as the melding of hacking skills, math and statistics knowledge, and substantive expertise. It’s the hot term for the combination of skills that more and more industries need.

With that in mind, IRE and the Center for Investigative Reporting data journalism team created a Kaggle competition that asked data scientists to look at campaign finance records. More than a dozen new ways of looking at the data came back, and with it, some insights into how journalists could learn from the various approaches. Chase Davis, who led the CIR team, talked about the results and provided his own code, slides and tipsheets from his four talks on GitHub.

One of the key tools for statistical analysis is R, and two of its most popular ambassadors gave hands-on demonstrations of how to use it. New York Times graphics editor Amanda Cox’s session showed attendees R’s power to generate maps from data that she had painstakingly (and considerately) cleaned beforehand. A sample of her more recent work can be found at NYTimes.com.

Hadley Wickham, statistician and author of several popular R libraries including ggplot2 and plyr, held a daylong workshop that delved into ways to visualize, clean, transform, and model data with R. For many in the course, it was an eye-opening introduction into how to use the tool, but more importantly, how to understand, doubt, and test datasets. Hadley has shared his detailed slides and code and Sisi Wei shared her class notes.

Making data journalism easier for everyone

tabletop-js-google-doc

There was much discussion of tweaking workflows to make the reporting process more data-journalism friendly. Last year, Balance Media and WNYC introduced Tabletop.js, which allows Google Spreadsheet data to power web interactives. This year, the Chicago Tribune news apps team introduced Tarbell, a Google Docs-driven CMS. Journalists Heather Billings (Chicago Tribune), Jacob Harris (The New York Times), and Al Shaw (ProPublica) spoke about this and other ways of getting news apps and the CMS to live together in their talk Infect the CMS.

Since 2010, NICAR has hosted a lightning talks session. Attendees get to pitch a five-minute talk, and the 10 most popular are presented. After last year’s “Cats Cats Cats” stunt by Aron Pilhofer of The New York Times, it was no wonder this year’s session was packed. This year’s standout moment was Ben Welsh’s five-minute rant (and yes, there’s some cursing) about the five ways coding like a web developer can make you a better investigative developer. The best part? Ben tells coders the five things need to learn from reporters. Even as the methods and tools change, tried and true reporting skills still matter.

Be excited and keep learning. Visit the complete roundup of NICAR13 tools, slides, and links and dig in.

Cartoon via xkcd.

Meet the first class: Six media startups get accelerated into Silicon Valley by Matter

Posted: 06 Mar 2013 06:00 AM PST

matterhq

Matter, the startup accelerator focused on media entrepreneurs, just launched its first class, which includes startups focused on interactive publishing, video analytics, and audio storytelling. The six companies selected for the accelerator are spending the next four months at Matter’s development space in San Francisco, honing their products and business plans. Drumroll, please:

The collection of early-to-mid-stage startups will embark on a bootcamp that mixes the atmosphere of Silicon Valley with the mindset and goals of the world of public media — the unusual mix of styles that makes the project so compelling. Each team will receive a $50,000 investment along with coaching from people in the worlds of media and tech, culminating in a final demo day for potential investors in June.

Matter CEO Corey Ford said the selected startups share a common mission based in informing society and engaging communities. “What we wanted to do was build a disruptive playground for media entrepreneurs to build a more connected and empowered society,” he said.

Matter is backed by funding from Knight Foundation and San Francisco’ KQED, both of whom get an equity stake in successful projects. More grounded in the world of venture capital than philanthropy, Matter occupies a different corner of the continuum of journalism funding techniques; the accelerator is focused on creating what Ford calls “feasible, viable, and desirable” products.

“If we want to drive innovation, I think having viability constraints is very important,” Ford told me. The startups at Matter, Ford said, have to push beyond good ideas that can support journalism and focus on creating plans that will make them work in the open market.

It’s like a Y Combinator for people with a public media mindset, in other words. Ford described the program as a series of four one-month sprints, where the teams will be in a constant state of refining and testing their products. The point of bringing together the six companies is to create a sense of community, where the teams can collaborate and learn from each other, Ford said.

Ford’s background is an appropriate mix of public media and Silicon Valley — he’s a former Frontline producer (an Emmy- and duPont -winning one at that) who has taught design thinking at Stanford’s Institute of Design, and he helped create Runway, an incubator program for Eric Schmidt’s Innovation Endeavors. Ford said the idea of design thinking — taking a human-centered approach to developing new products and ideas — is what can connect the worlds of Silicon Valley and media.

A lot has changed since the public media accelerator was first announced in December of 2011. Originally, it was going to be located in Cambridge and be operated as a project out of the Public Radio Exchange (PRX). Jake Shapiro, CEO of PRX and a partner in Matter, said the idea shifted from being a incubator for public media into a for-profit investment firm. Initially the public media accelerator was supposed to be a way to attract the kind of coders, developers, and entrepreneurial talent that was side-stepping the world of media for the tech sector. That’s still the goal, Shapiro said, but it works hand-in-hand with creating investment-ready businesses. “It’s a key part of this — the focus on changing media for good and innovating the industry,” Shapiro said.

Finding and maintaining a balance between the worlds of media and tech will be the key for Matter. Setting up shop on the West Coast means a lot of exposure to the world of Silicon Valley. That’ll be leavened by the resources of Knight and KQED, who can provide knowledge and access to the world of media, Shapiro said.

There is a lot riding on Matter’s first class, from the startups to funders like Knight and KQED. The accelerator represents a shift in media investment for both, with Knight continuing to experiment with tools beyond traditional grantmaking, and KQED looking for potential technology to put to use in its own network.

Ford said Matter represents a new model for investing in media innovation, which means they’ll be building a future not just for the startups, but also the program itself. But he emphasized that Matter is just one part of the ecosystem for developing and supporting new technology and business models in media. “I see everything as prototypes,” Ford said. “I see this as Prototype No. 1.”
di