« « Why technology matters: It’s about reporting

Collaborating with computers to parse “big data” » »

Data Delver: Ted Mellnik, Charlotte Observer database editor

Computer-assisted reporting is important because of its potential for reporting and analysis.  Visualization is important to present the information to readers. They both fall under the responsibilities of Ted Mellnik, database editor at the Charlotte Observer. His passion for data is as clear from a conversation with him as it is from his work.  Combine that with a keen understanding of the strengths of his colleagues, and you wind up with compelling stories and clean interfaces that get Charlotte citizens the information they need to know.

Mellnik is the first of my interviewees in a series I’m calling “Data Delvers,” where I’ll pass on summaries, quotes and audio clips from conversations with journalists using technology to find, analyze and convey data-driven stories and/or projects to the modern audience.

How the data journey began

Audio: How Mellnik got interested in CARMellnik’s first encounter with computer-assisted reporting.

- While the notion of the programmer-journalist is catching on in popularity now, Mellnik’s interest in computers began decades ago, as he started exploring programming on his own.

- At the same time, while he was a general assignment reporter at the Observer, he first encountered CAR when he observed a reporter using spreadsheets to analyze the 1983 election in which Harvey Gandt was elected Charlotte’s first black mayor.  Mellnik started to delve into that technology, realized it would be necessary to learn more about working with a relational database manager and it took off from there.

- “I was able to leverage an assignment to get more training,” he said.  When he was asked to cover the 1990 census, he told his editors that he needed to learn more about 9-track tapes and Unix to understand how the data was stored.  They obliged, and he was able to use the knowledge for that story, as well as others.

Taking a CAR story from data analysis to visualization

Audio: Mellnik’s collaboration with the graphics team at the Observer.Collaboration with the graphics team is essential.

- “Anytime you do a story, you have an interest in how it’s published,” he said.  “When working on a typical story, you have a photographer and a graphic artist.”  But, he said, data stories require some different skills.  Mellnik often has a general idea of how to display something, and talks to the graphics team to finesse the visual experience.

- After five years of working with data, Mellnik took on more of a specialist role and partnered with reporters on a variety of beats, taking on the data aspect of various stories.

- “It became easy to be the person who was going to look after graphics and the visualization part, it was kind of a logical division of duties,” said Mellnik.

Unemployment map: a case study (Find the piece here)

- Mellnik built this piece as an experiment intending to see how simple a map the Observer could create that would convey the most data possible.  It primarily uses Javascript to act on a cached layer that already exists on a map server.

- The map was designed to be simple at first glance, yet convey detailed information through user requests.  The charts contained in each pop-up are populated via the Google Charts API.  Mellnik points out that the current month (at the time the map was built) and that same month a year prior, are the only two values emphasized and displaying an exact numerical value.  This is because the comparison is most accurate when done between the same months a year apart, since the U.S. Bureau of Labor Statistics data is not seasonally adjusted on the county level, like it is for national and state-wide data.  Thus, changes may have more to do with other factors that are not relevant to the unemployment rate.

- The issue of the lack of the seasonally-adjusted data is not included in the map. According to Mellnik, this is because, “there is only one flavor of county-level data available, and the numbers are widely published…We tried to point people to what was the best comparison in the charts.”

Reaction to data work from colleagues and the public

- Mellnik said the Observer places high value on maps and data stories, in both their print and online editions.  For the recent presidential election, there were three or four maps in the paper the next morning, and another two or three the next day, he said.  “The reason was that the editor of the paper said we can tell this story through maps,” said Mellnik.  “It’s just a quick immediate way to convey patterns.”

- “I feel like I’m lucky, there’s a lot of demand for maps in print,” said Mellnik.  “Maps are also good interactive features, and people here place high value on that too.”

- Feedback from the community has been generally positive.  Sometimes Mellnik receives suggestions on tweaks for maps, which he likes to implement to “fine-tune the design.”

Other snippets of advice

Audio: The difference between mapping points and polygons.The difference between mapping points and polygons

- On learning SPSS vs. R: “Proprietary software is often easier to use and learn, and sometimes its features are half a step ahead of open source…you just have to make the best decision for what you think your situation is.”

- “Because the presentation side is so colorful and satisfying, don’t lose sight of the fact that you’re working as a reporter.  When you get data, it’s not always presentation appropriate, sometimes there are the stages of analysis, cleanup, exploration and summary before you can get to the presentation.”

- Mellnik said that data gathering is like any reporting, there’s a significantly larger amount of information behind a piece than what actually makes it into the final story.

« « Why technology matters: It’s about reporting

Collaborating with computers to parse “big data” » »