Just hours after two planes crashed into the World Trade Center, while all of us were still absorbing the immensity of the tragedy, Newsday reporters began writing profiles of the victims for the next day's paper. It has become the newspaper's forte, after winning a Pulitzer Prize for its coverage of the 1996 crash of TWA Flight 800.
Over the next several weeks, reporters gathered names of victims from many different sources: Web sites of companies who lost employees, press releases, relatives, airlines, the medical examiner's office, wire services, death notices, the police and New York State Supreme Court, which, at the request of relatives, issues death certificates for victims.
Reporters then turned to Nexis and AutoTrack to track down the victims' hometowns, ages and phone numbers.
A month after the disaster, we had collected the full names and hometowns of nearly 2,650 victims. The database, now even larger, is available online at www.newsday.com.
Profiling victims
The database proved useful for editors to keep track of which profiles had been written. But it also enabled three of Newsday's computer-assisted reporting specialists - Robert Fresco, Jo Craven McGinty and myself - to write a story showing just how far-reaching the acts of terrorism were.
Analyzing the database in Microsoft Access, Newsday's enterprise team wrote a story revealing that victims came from at least 41 states and the District of Columbia. Using ArcView, we created a map of the United States, highlighting how many victims came from each state. Another map showed the metropolitan area, with various sized dots indicating how many victims perished from each community.
The maps were powerful images, with all but nine states highlighted on the U.S. map and with dots small and large all over the place on the metropolitan area map. The map clearly showed a cluster of victims within easy commuting distance of Manhattan. Most of Long Island's victims, for instance, came from Nassau and Western Suffolk counties, each within an hour or less commute by train. (That's considered an easy commute in New York.)
We showed that the terrorists took the lives of people from more than 770 cities and towns - from Portland, Maine, to Honolulu, Hawaii.
We also produced a chart showing the number of victims in each age group, by decade. More than a third of the victims were in their 30s, and nearly half hadn't yet celebrated their 40th birthday.
Figuring percentages
We then determined the gender, based on the first name of the victim. The Census Bureau provides a database of male first names and female first names at www.census.gov/ftp/pub/genealogy/names/, including how frequently each name occurs for men and women. It's a little quirky. For instance, both the male and female databases list "James." But, only one-hundredth of 1 percent of women were listed as having the name "James," compared with 3.3 percent of men.
Of course, some names are both male and female, such as "Jamie" - 0.07 percent of men and 0.15 percent of women. When a name was listed in both the male and female databases, we used the frequencies to determine whether it was male, female or both. So, as expected, "James" turned out to be a male name while "Jamie" could be either. When calculating the percentage of male and female victims, we used only the names that were either male or female, excluding those that could be either.
We then supplemented the Census database with another database of names we found at www.incompetech.com/named/. The database is searchable online, and the Webmaster e-mailed us a link to download it.
Using these lookup tables, Newsday estimated that three-quarters of the victims were men. (OK, that's a lot of work for something seemed obvious given the profiles we had been doing. But it gave us an approximate percentage.)
We, of course, revealed the shortcomings of the analysis. Our database was skewed toward mostly male victims of the New York City police and fire departments, which released the names of virtually all their victims. Many of the companies that lost employees have not released lists.
Richard J. Dalton Jr. can be reached by e-mail at rdalton@newsday.com