San Francisco’s bus program Data from our set was correlated with homeless-count numbers reported to the Department of Housing and Urban Development. We grouped travelers by country and, using the average geolocation of destinations within each country, we calculated the distances from New York.ĭifference in mean income of origin and destination cities For the mean incomes in the US cities in our dataset, we scraped the US Census Factfinder, cross-referencing with the US Cities Database to ensure the income data was matched to the correct city. To produce per-capita homelessness rates, we used 2016 state homeless population estimates released by the Department of Housing and Urban Development and 2016 state population data from the US Census Bureau.įlights from New York Using Google Maps APIs (through the ggmap package in R) we attached a geolocation to all of the non-US trips from the New York dataset. US map We analyzed 21,000 journeys taken on the US mainland in 2011-16. We also noted idiosyncrasies in some programs, for instance in the small scheme in Humboldt County, California, where staff said tickets are in the main given to homeless people but others in need also receive them. A large portion of the data comes from either New York or San Francisco, and we bore this in mind when deriving insights from the set. These were further analyzed to understand trends in age, gender, race, and destinations, as well as changes over time, depending on which variables were available. If the destination name was still unclear and an internet search did not produce results, we deleted the row from the dataset.Īltogether we collated information on 34,000 individuals taking journeys. We confirmed each of these suggestions manually. Although the vast majority of data points had an exact match, when no match was found an algorithm was applied to weed out typing errors. The spreadsheets received from the 16 cities and counties required extensive cleaning and categorizing, and the US Cities Database was used to map each bus ticket to a destination based on census data. Complementing these were numerous interviews with program managers and homelessness experts across the US. In total these efforts yielded more than 35 narrative accounts of journeys, and we were also able to travel with two homeless people on buses. Two reporters also rode along with the Reno police department to meet travelers. Owing to reticence from officials, however, two San Francisco-based reporters stationed themselves outside the city’s Homeward Bound office several days a week over a period of a month in order to meet and chat with ticket applicants. In order to accompany a homeless person embarking on a bus journey, we requested introductions from city programs.
In addition, reporters in New York, Los Angeles, Portland, Key West and San Juan spent several days visiting shelters and homeless encampments. To interview people who had taken bus tickets, we called the hundreds of numbers we received, and for the 700 names for which we did not have numbers, we searched for contact details on the Nexis database and sent messages on social media. Most of the cities supplied spreadsheets documenting the destinations and dates for individual journeys, some provided the names of travelers and three gave their phone numbers. After producing a shortlist, we submitted several dozen public records requests to cities in order to obtain their data.
We also searched for references to bus programs in news archives from the early 1980s onwards. To determine which American cities have homeless bus programs, we contacted homelessness, housing or police departments in the 25 largest cities in the country.
Reporting on this project began in mid-2016.
Alastair Gee (homelessness editor), Julia Carrie Wong (reporter), Paul Lewis (editor), Adithya Sambamurthy (video editor), Charlotte Simmonds (copy editor)Ĭarla Green, Erin McCormick, Winston Ross, Thacher Schmid, Luis Trelles, Amanda Waldroupe, Joanna Waltersĭaniel Hollis, Sara Lafleur-Vetter, Michael Landsberg
0 kommentar(er)
