I get to work with all kinds of data at USA TODAY, but the book list has been a constant. When I arrived at USAT in 1997, one of the first projects I took on was to build and analyze an archive of the list to mark its fifth anniversary. Since then, as that archive grew to hold nearly 18 years of data, we’ve used it to anchor stories about authors and trends in publishing. We’re awfully proud of the list, and people in the publishing industry tell us it’s one of the most accurate accounts of Americans’ weekly reading habits.

Last year, we opened the archives up to developers via a Best-Selling Books API. This year, giving the list itself a facelift was the next logical step.

We were fortunate to assemble a crack team of designers, developers and product managers who, in a short time, conceptualized, designed, redesigned, and coded an entirely new collection of book-related pages for our site. What’s new:

– The list itself has an all-new design, including book covers.
– You can filter the list by genre or type. Handy if you enjoy a certain kind of book.
– Dig into the archives. Search by title, author or in the book’s brief description. You also can see entire lists from earlier weeks. Here, for example, is the first book list USA TODAY published, on Oct. 28, 1993.
– Clicking a book title takes you to its own page, which includes stats, reviews from Goodreads and latest Tweets about the title (more reader reviews to come). Here’s the page for To Kill a Mockingbird, which has been on the list for 755 weeks so far.
– Links to buy the title from some of readers’ favorite sources. And since our developers optimized the site for tablets and phones, you can find, buy and read a book quickly.

All this is part of an overall redesign of books coverage, which in turn is part of a larger site redesign.

My role in all this was to help the team understand the intricacies and particulars of the list and suggest the most suitable interactions it could offer. Coding-wise, I wrote some SQL to support the fetching and searching of the lists and the individual title’s archive data.

What a great time, a nice next chapter after finishing our work on the Census 2010 P.L. 94 release. More to come …

I can’t know for sure, but I suspect that most people who read my site are journalists or developers. Most traffic comes from links I post on Twitter or via search keywords that tend toward journalism, data, math and, lately, the Census.

Generally, you’re not an IE-centric crowd — just 12%. That’s lower than overall metrics, which tend to place Internet Explorer at anywhere from 40% or more of the overall market.

Oh, and the percent using IE6? Less than 0.4%.

Original post:

Stieg Larsson’s Millennium Trilogy grabbed the top three spots in USA TODAY’s annual list of top-selling books, reflecting a broader move by readers toward fiction this year. The list of 2010′s best-sellers was part of a package published today wrapping up the year’s trends as reflected in USA TODAY’s Best-Selling Books list.

Shepherding the books list is one of those tasks I spend considerable time on and, over the years, it has become one of my favorite opportunities for data journalism. The well never seems to run dry on ideas, and with 17 years of data in our archives, there are plenty of opportunities to see how annual moves on the list stack up against long-term trends. Along with Larsson’s success, our book team’s report on trends highlighted titles and authors reaching No. 1, from Nicholas Sparks to George W. Bush.

Much of our book list data is open for developers. Check out the API for details.

That happened last month, when the National Center for Health Statistics released data showing the number of births in the U.S. has dropped for two years in a row. One possible reason, the experts said, was the recession.

When it’s newsworthy, a yearly update to a longitudinal data set certainly is worth covering. But sometimes these basic demographics — including Census data — reveal even more when we take a long-term view.

For example, below are the annual number of births and deaths from 1933 to 2009 plotted in Many Eyes. Click the graphic to interact:

It’s simple — just two fever lines. But it’s chock full of  generational milestones that bear watching:

• The first baby boomers — those born in 1946 — turn 65 starting in January.
• The Gen Xers that follow have closed in on middle age. They range from the early 30s to mid-40s (in fact, Gen X poster boy Kurt Cobain would have turned 43 this year).
• Meanwhile, the first of the Millennials — the “echo boomers” whose numbers peaked in 1990 — are nearing 30.

Each generation brings a new sensibility to the stages of life, and the relative size and makeup of each one — not to mention its cultural context — gives journalists plenty of opportunity for storytelling. Two examples:

• Much has been written about the big bump of post-World War II babies marching closer to retirement (maybe), Social Security, and the years where health care becomes a major concern. But what about the inevitable? Notice that the number of deaths in the U.S. has plateaued at about 2.4 million a year. That won’t last long with Boomers heading into the years where death rates rise dramatically. How will 4 million deaths annually affect the funeral home business, the ability to buy a cemetary plot, and the overall industry around end-of-life care?
• Along with Gen X came the “baby bust,” the years of rapidly declining birth rates that led to all kinds of prognostications about the shrinking of America. That means our workforce now has a relative shortage of thirtysomethings. Does that mean more opportunity for Millennials to advance in the business world and less pressure for boomers to retire?

These sorts of trends are slow-burning, but they reflect data trends that exert hidden but massive force on our culture, much like the tides. The savvy data journalist keeps an eye on them not just for what they say this year but what they reveal over time.

Excel (and other spreadsheets such as the open source Calc) make sorting data easy. In fact, I often make sorting my first step when “interviewing” data because it quickly reveals high and low values and often highlights some that may seem questionable.

Let’s work through a simple sort in Excel. I’ll be using Excel 2007, but older versions have similar functions. Start by downloading the file “sorting.xls” and saving it to your computer. Open it and follow along:

1. We have a table of Census data from the 2006-2008 American Community Survey. It shows the median age of the population for each of 79 school districts in Virginia plus the state itself.

We want to know which district has the oldest and youngest populations. Let’s sort it!

2. Click once on one cell anywhere in the table. This will help Excel auto-discover your table in the next step.

3. On the Excel ribbon, select the “Data” tab and click “Sort.”

4. Two things happened. One, your entire table was selected (or highlighted). Two, a dialog box popped up to offer sorting options. Check off “My data has headers.” That will prevent your header row from getting sorted with the data, and it will add the three column names under the “Sort by” drop down.

5. Under “Sort by,” select “Median.” Under “Order,” select “Largest to Smallest.”

6. Click “OK.” Excel sorts your table, ranking the districts by median age — from highest to lowest. Your first few rows should look like this:

Now, we can do a quick scan and look for patterns. For example, several of the “oldest” counties are in southern Virginia, far away from the Northern Virginia economic engine. Meanwhile, the district with the lowest age is Harrisonburg City Public Schools — with a median age of a barely-legal 22.8. Could the fact that the city hosts two universities have something to do with that?

Good fodder for reporting, all made possible by a simple Excel sort.

A couple of tips and cautions:

– A good general practice is to work on a copy of your original data. Because things happen.

– Excel does best at sorting when your table has a header row and is not contiguous to any unrelated data, such as footnotes. Insert blank rows and columns between the data you want to sort and any information you want to keep separate.

– I recommend selecting only one cell in your table before selecting the “Sort” button. If you grab more than one, Excel may attempt to sort only those cells rather than the whole table. The 2007 version asks if you want to expand the selection, but older versions sometimes do not. This creates the possibility that only some of your data would get sorted, which is a nightmare. Always make sure your entire table gets selected!

– You can sort by more than one field. In Excel 2007, click “Add level” in the sort dialog.

Questions? Tips of your own? Add them below …

I’m giving a talk this week at the IRE Computer Assisted Reporting conference on how we acquired and analyzed the federal data that helped fuel the story.

Major stories in the series include:

– Schools in the dark about tainted lunches
– Why a recall of tainted beef didn’t include school lunches
– Fast-food standards for meat top those for school lunches
– 26,500 school cafeterias lack required inspections

Our series spurred congressional calls for reforms to USDA policies, and in February the agency announced tighter requirements on companies that supply food to the National School Lunch Program, including stricter testing of meat.

Here was my advice:

Focus on a few: Instead of holding building-wide Excel classes or database journalism seminars, start with just one or two reporters who show a combination of interest and decent technical smarts. That lets you go deep on a couple of beats rather than spread yourself thin. Also, success breeds success. Watching a few reporters land great stories will possibly spur interest from others.

Have the right goals: Goals like “publish one CAR story a week” miss the point. Better objectives are to have data-thinking ever present in the reporter’s mind, have the reporter well-versed in her beat’s data sources, and have the reporter develop basic data skills. From that, stories will flow.

Inventory data: Speaking of data sources, have each reporter you work with find out the sets of data local governments keep. File FOIA requests for table layouts and database schemas. Get the data, then study it. That will spur story ideas.

Crawl first, run later: All the hot talk in data journalism these days is on Web frameworks and visualizations, but there’s plenty of work for the beginner in the land of Excel and Access. Build those skills as a starting point.

Your thoughts? Add a comment below …

Also mentioned: The Washington Post’s investigation of the Metro Red Line crash and The New York Times’ series on toxic water.

From the item, by Mark Katches:

Just to be clear, this is by no means a comprehensive list. It represents only a small, informal survey about stories that some highly respected investigative journalists are buzzing about.

Indeed, every year I am amazed at the quality and depth of investigative reporting that American newsrooms continue to produce even as the industry fights hard times. It’s an honor to be mentioned in that company.

Let’s say six kids go on a field trip, ages 10, 11, 10, 9, 13 and 12. It’s easy to add the ages and divide by six to get the group’s average age:
 

(10 + 11 + 10 + 9 + 13 + 12) / 6 = 10.8

Because all the ages are close, the average of 10.8 gives us a good picture of the group as a whole. But averages are less helpful when the values are skewed toward one end or if they include outliers.

For example, what if we add a much older chaperone to our field trip? With ages of 10, 11, 10, 9, 13, 12 and 46, the average age of the group rises considerably:
 

(10 + 11 + 10 + 9 + 13 + 12 + 46) / 7 = 15.9

Now the mean is not an accurate representation. The outlier skews the average, and no journalist should feel comfortable reporting it.

This is where calculating a median is handy. The median is the midpoint in an ordered list of values — the point at which half the values are higher and half lower. If the median household income in East Middletownburg is $50,000, then half the households earn more and half less.

Using our field trip, we order the ages from lowest to highest:
 

9, 10, 10, 11, 12, 13, 46

The middle value is 11, and that’s the median. Half the values are higher, and half lower. If there had been an even number of values, we’d average the two middle values to find the median. For larger sets of numbers, you can use the MEDIAN function in Microsoft Excel.

Given this group, the median of 11 is a much better representation of the typical age than the average of 15.9. That’s what makes median such a useful statistical measure. Scan financial news, and you’ll see medians reported frequently. Reports on housing prices often use medians because a few sales of McMansions in a zip code that’s otherwise modest can make averages useless. Same for sports player salaries where one or two superstars can skew results.

A good test: calculate the average and the median for a group of values. If they’re close, then the group is probably normally distributed (the familiar bell curve), and the average is useful. If they’re far apart, then the values are not normally distributed and the median is the better representation.

In today’s story, reporters Blake Morrison and Peter Eisler worked with me to examine data on the number of schools in each state that met a federal requirement to have two cafeteria inspections annually. We found that in eight states, more than half of schools reporting failed to meet that standard in 2006-07 and 2007-08 school years.

Meanwhile, the series continues to draw attention on Capitol Hill. This week, Sen. Kirsten Gillibrand, D-N.Y., called on the federal government to increase its standards for meat used in school lunches and to cut contracts with companies that repeatedly did not meet standards.