But first, David Weathington wrote down the location - another spot where the often-unseen homeless of the suburbs spend their nights.
"If you're a homeless individual," he said, "you are going to find any place where people aren't going to find you, where people are going to overlook."
Last Wednesday, Weathington, along with about 90 others in Chester County and thousands more across the country, went searching for those overlooked spots as part of the annual Point-in-Time count, a 24-hour census that provides a complex, albeit imperfect, image of homelessness in America on a single night.
The count, coordinated by the U.S. Department of Housing and Urban Development, starts with homeless people staying in shelters or transitional housing. Then, volunteers such as Weathington, an employee with the Chester County Department of Community Development, scour their communities for the outliers, those who - because of lack of shelter beds or an aversion to taking one that is open - are spending the night of the census on the streets.
In Chester County, where the temperature dipped to 3 degrees by the time the count ended at 2:30 a.m., nine homeless people were found out in the cold: four in an encampment in the woods in Kennett Square; two curled up in their cars in Downingtown; and three across West Chester borough.
One of the three was found sleeping in a stairwell of a parking garage, the same place he had been found during last year's count. Weathington said the man refused help. But when Weathington checked the next morning, the winter coat and box of food the group left had been taken.
That was one success on a night that had its frustrations, as Weathington's teams - one that worked from 10 p.m. to midnight and a second that set out from 12:30 to 2:30 a.m. - scoured West Chester and came up mostly empty-handed.
He wishes that meant the community's homeless had all found refuge in shelters or with friends. It's more likely, Weathington said, that some of those most opposed to seeking help were able to avoid being seen, because they knew the count was taking place or they saw the group coming.
Methods are imperfect
While most homeless on the streets in West Chester last Wednesday night traveled alone, carefully traversing the icy sidewalks with their heads down, Weathington and his five volunteers moved in a pack, their eyes wide and their flashlights scanning side to side.
They slowed in front of cars to look for seats leaned back. They inched forward in the narrow alleys between rowhouses. They split into groups in Everhart Park, shining flashlights beneath evergreens and under the deck of a blanketed gazebo as light from the streetlamps drifted across the hardened snow and illuminated their footprints.
Mostly, they just walked, finding no one.
Few will say the Point-in-Time count's methods are perfect - with thousands of volunteers canvas their communities at various hours of the day, each using unique approaches that can change year by year. And advocates for the homeless are divided over whether the resulting numbers can be trusted.
Critics say they can't, and they worry that the 4 percent national decrease in homelessness seen in last year's count gives the false sense that a serious problem is subsiding. Maria Foscarinis, the executive director of the National Law Center on Homelessness and Poverty, said the system also led to the federal government's using inaccurate numbers to set priorities and funding.
The problem with the Point-in-Time count goes deeper than a bad-data set, she said: By counting in a single night those living in shelters or on the streets, HUD is disconnecting homelessness from its causes.
"In reality, homelessness happens over time," Foscarinis said. "It doesn't happen during a few hours on a single night. It happens in steps. And it's important to look at the sequence of events, because that takes us back to the cause."
Others concede the count is imperfect but is still useful in tracking a community that is inherently hard to pin down.
Nan Roman, president and CEO of the National Alliance to End Homelessness, said Point-in-Time offers a picture of the problem, while a deep view is given through data sent to HUD from service providers who work one-on-one throughout the year with homeless people.
"Sometimes," Roman said, "I feel like people are criticizing the PIT counts for not being something they were never intended to be, which is a much more detailed inventory of all the problems that surround all the people who are homeless."
Weathington said the count was useful to his office not just because the numbers become the basis for program planning and grant applications, but also because, each year, the process turns up a handful of new encampments and hideaways. Even if no one is there on the night of the count, outreach organizations can come back in the future.
Not being found
There's one spot Weathington knows should be revisited, where he thinks he and his team last Wednesday narrowly missed connecting with a homeless person who was set on not being found.
In a cluttered storage space near the entrance gate of a parking garage in West Chester, the group found a tightly bound plastic bag among the mess of pallets and cardboard boxes. Unable to undo the knot, one volunteer ripped a slit.
Inside, the group found a silver radio and a thin fleece blanket nearly faded of its pastel print.
The owner wasn't there. So Weathington put the bag back in its place and left, planning to return with the second shift of volunteers.
When he did, the bag was gone.