Tips for working with Get it Done data

Use these tips to perform a more accurate and efficient analysis of Get it Done data.

articles Beginner 7 min read

Get it Done is the City program that intakes reports of non-emergency problems from the public and notifies City departments and external agencies responsible for fixing the problem. People can submit a report through the mobile app or the website. The program has received and processed more than 3 million reports since it launched in 2016.

Tip 1: Answer the right questions

So many reports can give some insight into the problems that exist across the City, but this dataset shouldn’t be treated as a comprehensive list of problems. Some areas of the City may experience under or overreporting, and the factors that contribute to that have not been studied. For example, there could be a surge of reports about broken streetlights in a particular area. This could be because there was an increase in broken streetlights OR it could be because people in the area decided to send in more reports, possibly because they recently learned about the Get it Done program or a news story brought attention to the issue and influenced people to report.

Closed reports also do not represent a comprehensive list of work City departments performed. City departments plan work in other ways, for example by creating holistic asset management plans that rely on available funding. While many reports do result in work being completed by City departments, that’s not always the case. Sometimes, the problem has already been reported, the problem is the responsibility of another government agency or private business, or the problem cannot be found or verified. A status of “closed” does not necessarily mean work was performed (more on that below).

Get it Done is meant to make it easier for people to connect with the City; it’s a customer service program, so the data are best used to for analyses related to customer service. Some questions you could answer with this data include:

  • How many reports did the City receive?
  • Are there any geographic trends in the reports?
  • What types of reports were most popular?
  • How long does it take for the City to respond to reports, on average? Does that change depending on the report type? (“Respond” here means communicate back to a customer inquiry, not necessarily fix something)

Tip 2: Download the correct data

All reports are published here on this portal in the Reports of non-emergency problems submitted to Get It Done dataset. All 3 million+ reports are available in one downloadable file, but that is too large to be analyzed with Excel. Instead, choose from among pre-filtered views of the data to make analysis easier! Below is an overview of the available views.

Filtered by status and year

Status refers to whether a report is open or closed. There are two status files, one for each status overall, and the closed reports are further divided by year closed.

Filtered by type

There are more than 1,000 distinct types of reports in the data, but a few of them have emerged as some of the most popular problems to report. These are published in their own files.

  • Graffiti
  • Illegal Dumping
  • Potholes
  • 72-hour parking violations

Filtered by Council District

Because reports are geolocated, they can be sorted into one of the City’s 9 Council Districts. This is the only variation with a geographic filter available, although reports do include latitude/longitude coordinate pairs that can be mapped and further analyzed.

Tip 3: Learn what the columns mean

Each dataset on this portal comes with a data dictionary that describes the columns in the data. This should help clarify how to use each column in an analysis. Below are some of the columns to pay particular attention to for the most accurate analysis possible.

Request IDs

The unique identifier for a single report is the service_request_id. Some reports are determined to be duplicates. That giant pothole on a major thoroughfare was probably noticed by lots of people, and there’s a chance more than one of them reported it to the City. In cases like that, one of the reports is considered the “parent”. All other duplicate reports will have the service_request_parent_id field populated with the service_request_id of the parent. For a question like, “how many times were potholes reported”, count all reports. For questions like, “how many potholes were reported”, it might be important to count only those reports that aren’t duplicates. In that case, filter the data for reports with a blank or null service_request_parent_id.

City Departments

case_record_type in the data contains values that indicate the City group responsible for a problem. These don’t always map with a city department 1:1, but they offer an idea of which department might be responsible. It’s helpful to review the current organizational chart and mission/vision/values for each department. Keep in mind that the City reorganizes from time-to-time, but the data values for historic records won’t change. For example, TSW refers to a former department that combined Stormwater and Transportation Departments. These departments are now split. Another caveat is that some reports are handled by multiple departments. Encampment reports are routed both to Environmental Services Department for cleanup and to Neighborhood Policing for outreach and other services for those experiencing homelessness. And finally, some reports are referred to external agencies. Those gain a “Referred” status and a value of the referred agency in the referred column.

Status

Reports will have one of four statuses: Closed, Referred, In Process, and New. New and In Process are both statuses that are considered active and/or open; these reports are still being reviewed, routed, and worked on. Closed and Referred are statuses that are assigned to inactive and/or finished reports. At this point, the City has concluded its review and routing; however, it does not necessarily mean work to fix the problem was performed.

A referred case will go to one of two places: an external agency, such as San Diego Gas & Electric or the Metropolitan Transit System, or a City department that reviews and routes its reports in a different software application than the one that powers Get it Done. Either way, the Get it Done program no longer has visibility into what is happening with the report and unfortunately cannot give any more updates back to the customer who reported the problem. For this reason, the report is deactivated in the Get it Done software application with a “Referred” status. The team continually enhances the software with the eventual goal of adding every City department that provides public-facing services. To whom a report is referred is contained in the referred column.

A closed case was fully routed within the Get it Done software application. The customer who made the report will receive some details about what happened when the City assessed the problem they reported. These details are outside of the scope of the current dataset and are not included, but in most cases, the City department closes a report after completing work to fix the problem. Sometimes the City closes a report for other reasons, for example, staff could not find the reported problem or the problem will be addressed in the future as part of a holistic asset management plan.

Type of report

For analytics regarding the types of problems that get reported, the service_name and service_name_detail columns will be helpful. The service_name column is a top-level category, with service_name_detail giving, well, more details. The dataset does contain more than 1,000 combinations of the two columns. This is because departments vary in the number and makeup of their work groups and need different levels of detail to properly route reports to their groups. Unless the analysis is looking at a very specific problem, it should be sufficient to group reports by the top-level service_name.

Speaking of specific problems, another way to understand the type of problem is to hunt for keywords in the public_description field. A member of the public may have a hard time knowing which top-level category to choose when submitting a report, so some categories may go undetected in the “Other” group or be mis-categorized.

Tip 4: Skip the analysis and check out one of our dashboards

These tips should make any analysis easier, but another option is to check out exising City dashboards. Check out the year-over-year count of reports coming in, grouped by current status here.