Driven Into Debt

Donate

The Ticket TrapAbout

The Ticket Trap: About Our Data

Details behind our database on ticketing patterns across Chicago’s 50 wards.

by David Eads and Melissa Sanchez

ProPublica Illinois’ interactive database The Ticket Trap reveals disparities in ticketing, debt accumulation and appeal rates across the city’s 50 wards. It relies on data obtained from the city of Chicago in partnership with WBEZ; our organizations have worked together on and published a series of stories that examine how uneven ticketing practices and increased fines have led to more debt for the city’s poorest residents, prompting thousands of bankruptcies.

Data sources

Most of the data used for this interactive is derived from a database of more than 54 million parking, standing and vehicle compliance tickets issued in the city of Chicago between Jan. 1, 1996, and May 14, 2018. We obtained the data through public records requests to Chicago’s Finance Department, which administers much of the city’s ticketing and debt collection operations. Citations from red-light and speed cameras are not included in the database.

Because the data set is so large, city officials released it in two waves:

  • Information on tickets issued between Jan. 1, 1996, and Dec. 31, 2006. This data was last updated on Aug. 27, 2018.
  • Information on tickets issued since Jan. 1, 2007. This data was last updated on May 14, 2018.

The data set that includes the most recent tickets is available via the ProPublica Data Store; we will update it to add the older records soon.

We obtained information about the aldermen, their office phone numbers and email addresses from the City Clerk of Chicago. After checking the clerk’s data independently, we changed three email addresses and added a tilde to the last name of one alderman.

We obtained 2015 ward boundaries from the city’s Data Portal.

Our interactive database also uses data from the 2013-2017 American Community Survey to provide some demographic context for each ward.

Geocoding Addresses

Every record included where the ticket was issued. We edited those addresses to the block level; for example, 123 N. State St. became 100 N. State St. We then ran all of those addresses through the Geocodio geocoding service to determine the latitude and longitude of each citation and to place them in wards, using the latest boundaries. To test our accuracy, we compared our results with those from another service, SmartyStreets, obtained by local civic hacker Matt Chapman.

About 12 percent of the tickets could not be properly geocoded because of errors in how the addresses were spelled or entered into the city database. We found that about 6 percent of the remaining tickets may have been geocoded incorrectly and assigned to the wrong ward. We are currently studying how much actual variability these errors introduce in the data we display. As a result, ward-level summaries are conservative estimates that err on the side of undercounting the extent and effects of ticketing in many categories, including how many tickets have been issued and how much money is owed to the city.

Because of the possibility that messy data led to tickets being geocoded into the wrong wards, the maps should not be considered perfect representations of the geographic distribution of tickets.

Citywide Variables

We calculated the total number of tickets, total amount paid and outstanding debt using ticket records from Jan. 1, 1996, to May 14, 2018. We calculated the portion of tickets issued by police using records from Jan. 1, 2013, to Dec. 31, 2017.

Variables by Ward

We derived a number of variables from the ticket data for each ward and ranked the wards against one another. In most cases, we used data from tickets issued between Jan. 1, 2013, and Dec. 31, 2017, to display a recent trend that overlapped with the period covered in the American Community Survey. For our variables on payment and debt, however, we used the data on all tickets dating to 1996. We did this because the ticket debt is still on the books, potentially threatening motorists’ vehicles, licenses and livelihoods.

  • Amount due: The sum of all tickets and late penalties due in a given ward from citations issued between Jan. 1, 1996, and May 14, 2018. A small number of tickets have a negative amount due. We take this field at face value: The negative amounts are included in our summations.
  • Percent paid: The portion of tickets marked as “paid” in a given ward, from citations issued between Jan. 1, 1996, and May 14, 2018.
  • Amount paid: The amount of money paid, including any late penalties, for tickets issued between Jan. 1, 1996, and May 14, 2018.
  • Ticket count: The number of tickets issued between Jan. 1, 2013, and Dec. 31, 2017.
  • Portion issued by police: The portion of tickets written by police officers, and not city parking enforcement aides, private contractors or others who can also issue vehicle citations, between Jan. 1, 2013, and Dec. 31, 2017. Citywide, police issued 39 percent of all tickets during that period.
  • Top types of tickets: This shows the five most common violations between Jan. 1, 2013, and Dec. 31, 2017. We’ve also included the cost of each ticket, without late penalties (which can add as much as 122 percent to the original cost). In a few cases where the ticket price increased during that time, we displayed the most recent price.
  • Average ticket price: This shows the average cost of tickets issued in a given ward between Jan. 1, 2013, and Dec. 31, 2017.
  • Associated with notice of vehicle or license seizure: The percentage of tickets issued between Jan. 1, 2013, and Dec. 31, 2017, that were tied to a notice from the city of a pending vehicle seizure or license suspension on the date the data was extracted. Two unpaid tickets can trigger a boot on a vehicle. The city can initiate license suspensions after motorists accumulate 10 unpaid parking tickets or five unpaid traffic camera tickets.
  • Included in a bankruptcy: The portion of tickets issued between Jan. 1, 2013, and Dec. 31, 2017, that were tied to a bankruptcy on the date the data was extracted. Thousands of motorists with ticket debt file for bankruptcy each year to get legal protection that allows them to hold on to their driver’s licenses and vehicles while their cases are active. Most bankruptcies end in dismissals, meaning without debt relief; when that happens, the city can once again go after motorists for unpaid tickets.
  • Percent contested: The portion of tickets issued between Jan. 1, 2013, and Dec. 31, 2017, that were appealed through the city’s administrative hearings system.
  • Percent found not liable after appeal: The portion of tickets issued between Jan. 1, 2013, and Dec. 31, 2017, that were marked “not liable,” meaning they were thrown out, after a motorist appealed at an administrative hearing.

Limitations

Because the data represents a snapshot of a point in time, we can’t determine trends in time in key categories such as bankruptcy or debt. We only know what those variables looked like on the day the data was exported.

We also know that most Chapter 13 bankruptcies that include ticket debt end in dismissals, meaning with no debt relief, and that motorists often file for bankruptcy multiple times. This means there are significantly more tickets that, at some point in history, were associated with a bankruptcy. The rate we show, however, is only of those tickets tied to an active bankruptcy on the day the data was extracted.

Finally, the data is on tickets, not drivers. We don’t know the race or ethnicity of the individuals who received the tickets, but Census data indicates the racial/ethnic makeup of the ward. For high-traffic areas like the 42nd Ward, the population of the ward and the population of drivers in the ward are almost certainly quite different. For residential areas, they are more likely to be similar.

Bulletproofing

ProPublica computational journalist Jeff Kao independently bulletproofed the process that created the data set we used by writing a script that ran the same filters and summaries used in the initial analysis, then comparing and reconciling the two sets of results.

Acknowledgements

Thanks to Chicago demographer Rob Paral for helping us understand some of the caveats with the American Community Survey data, and to Matt Chapman for sharing tips on how to geocode the tickets. We are grateful to many others for their valuable insight as we developed this interactive, but especially our engagement reporting fellow, Helga Salinas; WBEZ’s digital editor, Elliott Ramos; Kulsum Ameji and Martin Martinez from LAF, and our good friends at Chicago Hack Night and City Bureau’s Public Newsroom.

Have tips? Find errors? Let us know.

Are you looking at our database and have some information you want to share? Are you finding inaccuracies? We want to hear your thoughts. Please send an email to [email protected].