|Stories, projects, web pages at The Charlotte Observer
Cradle of Secrets
This six-part series tells how a state medical examiner's misuse of the term
SIDS stifled criminal investigations of suspicious infant deaths.
The newspaper examined death records and autopsies for more than 550 N.C. infants
labeled SIDS, and found that 69 percent died while they were exposed to
documented, preventable risks of suffocation. SIDS deaths are not fully investigated
by police or prosecuted because SIDS is an official finding
that a death was unpreventable. I organized the data and analysis.
The weekly map opens with the latest week of selected crime data from Charlotte-Mecklenburg
Police. It's made up of point layers of crimes stored in .json files
and a street background layer from Google.
The retrieval of details uses ASP.NET and SQL Server. The custom map page lets you draw on the
Google base map to outline a neighborhood, then retrieves up to a year of crime data for that custom neighborhood.
Neighborhoods That Win and Lose with Outdated Tax Data
I used public property sales and tax appraisal records to highlight neighborhood
disparities in property tax burden due to outdated assessments.
For a sample of 16 neighborhoods, this study estimated
how property taxes for 2008 would have changed if assessments had been consistent countywide.
Although theoretical, this scenario benefits from using established spending and tax levels, thus
eliminating much of the uncertainty that would cloud a forward-looking analysis.
|Nov. 30, 2008
For election night, I wrote scripts to download from state sites XML precinct results files
for counties in the Charlotte region and match the data with information compiled from
voter registration data and precinct maps.
I did similar work with county and state results files from AP. The resulting
maps and analysis went to newsroom intranet pages. I contributed to six county and precinct maps published in the
newspaper Nov. 6, and I supplied XML data files to drive interactive online maps created by
graphic artists in Flash.
|Nov. 5, 2008
|Nov. 6, 2008
|Nov. 7, 2008
The Nov. 6 regional precinct map and analysis package showed how Obama's support tracked nonwhite
voter registration, but also pointed to precincts where his support was significantly greater than
racial composition would have predicted. An interactive version of the map was available on the newsroom
intranet for reporters.
The Nov. 7 national map showed which counties were carried by McCain and by Obama, and which had switched
party since the presidential election in 2004.
Foreclosure's Tax Bite
Part of the property tax burden on foreclosure-plagued neighborhoods is expected to shift to healthier
ones after the next county property revaluation. I analyzed home sale data and county property records
to make detailed maps of price trends since 2003. Summary maps are based on a grid to avoid shortcomings
of outdated census geography. An interactive map online lets readers explore
31,000 benchmark home sales.
|April 27, 2008
New Suburbs in Fast Decay
Foreclosures lead to vacancies and crime in new suburban neighborhoods.
The neighborhoods once held promise for thousands of Charlotte
families, but now struggle with crime, blight and falling home values.
I used a kernel density analysis to map foreclosure concentrations and
compare them with crime reports. Neighborhood vacancies and rentals
were estimated by comparing addresses in public property and tax records.
|Dec. 9, 2007
Some Serious Waiting
Iraq and Afghanistan war veterans suffering traumatic brain injury, grave
wounds or serious illnesses often wait longer for outpatient appointments than
the 30-day VA standard, according to an Observer analysis of two internal VA
|Oct. 21, 2007
Sold a Nightmare
This four-day series tells of Beazer Homes' aggressive marketing to low-income buyers in the Southern
Chase neighborhood, and how the
company sometimes sold to people who couldn't afford to buy homes. Among major home builders, Beazer produced
the most area starter-home neighborhoods with high foreclosure rates. Local and state officials don't keep track.
The original presentation included an interactive map of the Southern Chase neighborhood.
The N.C. legislature has since enacted laws requiring
public records to name the person who originates each mortgage loan, and defining a new crime of mortgage
fraud. A Pulitzer finalist.
On the shores of Lake Erie, the city of Buffalo and other communities have withered.
Meanwhile, since 2000, Charlotte has added 46,000 residents.
An analysis of data from the Internal Revenue Service
shows a significant chunk of upstate New York's population has moved to the Charlotte region.
It shows that Mecklenburg County is the No. 1 out-of-state destination for people leaving Erie County,
home of Buffalo. It's the No. 2 out-of-state destination from Monroe County, home to nearby Rochester.
|Jan. 21, 2007
Foreclosing on the American Dream
This series looks into the rapidly rising numbers of home foreclosures,
and the effects on neighborhoods where failed home loans have concentrated since the advent of
easy credit by government and lenders.
The neighborhoods are often new subdivisions priced for first-time buyers.
But instead of building wealth, the buyers often lost their homes and damaged their credit.
Neighbors who pay their mortgages on time get hurt, too, because concentrated foreclosures can
depress home resale values.
The story was based on newly available data pinpointing foreclosed homes. Day 1 included a full-page graphic
mapping 4,900 foreclosed homes. The online interactive version of the map
has been updated regularly.
Following Hurricane Katrina, I created an Internet page that allowed reporters to compile, search and share a database
of information about victims as they were identified. When a reporter obtained a list of locations (no names) where victims were found,
I mapped the addresses and created an overlay with U.S. Census tract data.
These data sources figured in two Knight Ridder Newspapers stories and a map graphic that moved Dec. 29, 2005.
|Jan. 1, 2006
One story reported that construction defects and other human error may be responsible for nearly 600 people who died after the height
of the storm when canal floodwalls collapsed. Another reported that contrary to perceptions,
poor and black residents of New Orleans were not over-represented among Hurricane Katrina victims.
Published pages are from The Miami Herald.
The series was the first large-scale analysis of
the new data on home loan pricing. I compiled a database of 2.2 million
loan application records from data obtained from 25 of the nation's
largest lenders. Major finding: Blacks were four times as likely as whites
to get high interest rates on home purchase loans.
Hard Truth in Lending
Many questions about the data were answered by looking at
rates of denial and of high-rate loans by race and income of borrower, by lenders and by
type of neighborhoods, based on census tract data. I created a web page to explore
This Knight Ridder investigation found that injured soldiers who seek disability from the
U.S. Department of Veterans Affairs face lengthy delays and inconsistent rulings.
Discharged and Dishonored
I created a web page to explore data comparing disability ratings vets
received for selected ailments from regional VA offices across the country. Each segment
on the bar graphs includes a mouse rollover that displays its data.
I created two national county-by-county maps showing results of the 2004 presidential election. One was an interactive map online
that went live early on Nov. 3. The other map went into a full-page graphic published the following Sunday; the graphic
won an award of excellence from the Society for News Design.
2004 Presidential Election Map
The first series showed that some N.C. judges acquit a large majority of the DWI defendants who go to trial,
even though the defendants tested over the legal blood-alcohol limit of 0.08%. One Charlotte judge
tried 50 DWI suspects who had tested over the legal limit. He acquitted them all.
Subsequent series focused on DWI dismissals, lack of punishment for underage drinking and driving, and the relationship
between judges in three coastal counties and the lawyers that almost never lose DWI cases brought before them.
I began analysis for these stories by merging court case data with data on alcohol test results.
Forrester Research had made news with an estimate that 3.4 million U.S. jobs in 175 service occupations would move overseas by 2015.
I applied that analysis to detailed government occupation and earnings data to produce state and metro estimates of potential losses
in jobs and earnings. The story and graphics described national patterns, and reported that the Carolinas
could lose 124,000 jobs and more than $5 billion in wages.
The Rush Overseas: There Go Our Computer Jobs
Following a heavy, late-winter snow storm, Charlotte city officials decided to abandon their let-it-melt policy for
neighborhood streets and sent out their snow plows. A reporter requested a list of streets serviced, and I matched
the list against streets on a city streets database. The resulting map showed that some snow plows drove miles
past moderate income neighborhoods to make sure streets were cleared in some well-to-do
areas. The random nature of the patterns raised questions about fairness.
This series detailed the role of faulty maintenance in recent airline crashes, including a January 2003
crash in Charlotte. Data sources included National Transportation Safety Board data and reports on U.S. fatal air carrier accidents,
Federal Aviation Administration data on inspections of airlines and repair stations and enforcement actions related to
maintenance, and Bureau of Transportation Statistics data on airline spending.
Are Planes We Fly More at Risk?
In early March, the Census Bureau released, under a three-day embargo, a Census data file on county-to-county commuting. That afternoon,
I published an intranet Web page that enabled reporters across Knight Ridder to explore this data, the changes since 1990 and maps
showing in-flow and out-flow patterns for every U.S. county. A main package ran in print on release day, and the map Web page was
published on the Internet.
Death at the Track was a 16-page section and a web application that documented 260 deaths in U.S. auto racing since 1990,
and the sport’s shortcomings in addressing safety. The project involved 16 reporters, nine editors, four news researchers,
and four photo and graphics specialists. They collaborated in large part through intranet pages I created to compile, report and
search a database of deaths. Death list copy for publication was written and initially edited in web pages,
keeping the material in a convenient, multi-user environment until late in the production cycle.
Death at the Track
After a parent sued Charlotte-Mecklenburg schools to end three decades of race-based school assignments,
I used computer mapping and analysis software, and student and Census data
to design a school system. For the first time, a newspaper
was able to show readers what schools would look like under a system of neighborhood assignments by mapping
where more than 90,000 children live and where they would attend school.
Among findings in the January 1999 series: Neighborhood schools would create a group of super-poor schools, concentrated in the
predominantly black inner-city. More than one-third of the county’s children would attend severely crowded schools.
And crowding would hit black communities hardest.
The project was described in a chapter of
Mapping the News: Case Studies in GIS and Journalism, by David Herzog.
- “Grave Secrets,” a 2001 series on the failure of N.C.’s medical examiner system to investigate suspicious deaths. As a result, the medical examiner’s office in Charlotte, one of the state’s most overworked, obtained an additional investigator and pathologist; and a state study commission proposed system reforms. I analyzed data on deaths and investigations.
- “Broken Trust,” a 2000 series that documented 33 deaths in mental health facilities that were not reported to or investigated by state officials. In response, the General Assembly mandated the reporting of such deaths, and the state increased spending for investigations. I helped scour death data for unreported deaths.
- A 2000 study of criminal prosecution in Mecklenburg County for the five-part series "Doing the Crime but not the Time,” It reported that violent criminals in Charlotte were only half as likely to go to prison as elsewhere, and those criminals often went on to victimize others. Response included additional fund for the district attorney’s office.
- “Uncertain Justice: The Death Penalty on Trial.” This 2000 series reported racial differences in the use of the death penalty, and weaknesses of the death penalty system. I analyzed data on convictions and sentencing.
- “Home Buyer Beware,” a 1999 series on home construction flaws and failings of the building inspection system. I analyzed data on building inspections to document how little time inspectors spent on each home. As a result, Mecklenburg County increased penalties for construction flaws by builders.
- "Slighted in Sports", a 1997 package on the disparity in men's and women's sports. I collected and analyzed data on college sports scholarships, spending and participation the first year it was available;
- "Paid in Pain," a six-part series in 1996 on the failings of N.C.'s workers compensation system. The state responded by speeding claims processing for workers who suffer crippling on-the-job injuries. I analyzed workplace injury data to document injuries and delays.
- “Race Against Death,” a 1996 series that reported how patients were dying because Mecklenburg’s understaffed and overwhelmed emergency ambulance service was sometimes slow in getting to emergencies. I analyzed EMS call data for performance. In response, the service was reorganized and expanded.
- “Taking Back Our Neighborhoods.” I mapped the locations of 25,000 violent crimes and
analyzed neighborhood crime patterns. The 1994 series was a Pulitzer finalist.
Work from The Washington Post