Tenderloin Emergency: Where is the Data to Measure Impact?

Tenderloin Emergency Dashboard (as of 2/6/2022)


Overview

The City’s Tenderloin Emergency Plan is a long overdue response to the devastating crisis in the Tenderloin. It is vitally important that the City’s initiative succeeds, not just for the sake of the people suffering in the Tenderloin, but also for the benefit of the entire city. Homelessness and deteriorating street conditions affect all of our neighborhoods, not just the Tenderloin.


The purpose of the RescueSF Tenderloin Emergency Dashboard is to evaluate the City’s progress in achieving the priorities of the Tenderloin Emergency Plan. Instead of presenting all of the metrics from the City’s weekly reports, this dashboard includes only those metrics that most clearly relate to the overall priorities. This dashboard explicitly ties reported metrics to the relevant plan objectives, and the dashboard presents data in graphs to make interpretation easier. The analysis also points out the limitations in the data. All graphs and analysis have been prepared by RescueSF.


The City has been implementing the emergency plan since the end of December. Based on the progress so far, we can make some initial observations.


  • One of the main objectives of the Tenderloin initiative is to connect people to supportive services. Toward that end, the City has had 5,631 meaningful engagements, defined as conversations in which a staff member discussed the details and eligibility of a service with a guest, but the guest did not choose to connect with the service. The City has made 2,309 referrals, defined as a longer conversation that leads to a guest stating an intention to access the service.

  • The City has helped 180 unhoused people leave the streets for shelter. The shortage of shelter capacity is preventing the City from moving faster.

  • The arrest data does not show a surge of police activity in the Tenderloin.

  • The data generally provides limited insight into the City’s progress on achieving its objectives. Like most government data, the weekly reports reveal what the City has been doing, but the data doesn’t measure whether the City is making an impact. We know the City has been busy, not whether it is succeeding. Going forward, we recommend revising the weekly reports to include more direct measures of the plan’s priorities.


Background

On December 14, Mayor Breed announced the Tenderloin Emergency Plan. The plan identified eight priorities. In the first operational progress report, with reporting as of December 26, the City identified seven priority areas that had been developed with input from community stakeholders:


1. Drug dealing and violent crime

2. Open-air drug use

3. Lack of shelter and drop-in resources

4. Lack of safe passage and accessibility

5. Waste and debris

6. High levels of 911 medical calls

7. Illegal vending


The Department of Emergency Management (“DEM”), the agency tasked with coordinating the multi-agency response, issues weekly operational reports on the Tenderloin initiative. The report contains a list of data on more than 120 metrics from different participating agencies.


The weekly reports are difficult to interpret. Sifting through more than 120 separate metrics is overwhelming, and the reports do not highlight which ones are more important than others. While the metrics are organized based on the reporting agency, the report does not link the metrics to the plan’s overall priorities, listed above. As a result, the reports do not clearly show whether the City is making progress on the main priorities. Moreover, each weekly report contains the data on that week’s operations without indicating how the results compare to what has previously been achieved.


The rest of this document addresses each of the seven priorities from the Tenderloin Emergency Plan and considers the most relevant data from DEM’s weekly reports.


1. Drug dealing and violent crime

Figure 1 presents data on calls for service. This data has been included Since January 23. Over the past three weeks, calls for service have been relatively flat. It is not possible to reach any conclusions about the incidence of crime based on such a short period of reporting. A more glaring weakness is that these figures about general service calls do not provide insight into violent crime, which is the real focus of the first priority.



Figure 2 shows changes in felony and misdemeanor arrests. Weekly felony arrests started the year at 32, rose to over 40, and then fell back to 35 in the first week of February. Similar to the data in Figure 1, the arrest data in Figure 2 is overly broad, covering both violent and property crime, and does not directly measure progress on reducing violent crime.



Figure 3 shows weekly arrests for drug sales or possession for sale. Despite seeing some fluctuations, weekly drug sale and possession for sale arrests have been relatively stable, with 14 in the first week of January and 12 in the first week of February. Since the start of January, the police have made 76 arrests for drug sales or possession with the intent to sell.



Figure 4 presents data on the amount of drugs seized – both total narcotics and the broken out figures for fentanyl. The data fluctuates widely over the course of the month. Did the police seize a lot of drugs, or a relatively small amount? Without knowing the total amount of drugs flowing through the Tenderloin during a week, it is not possible to reach any conclusions from this data. Commentary from the police could help make more sense of these results.



2. Open-air drug use

As noted in the previous section, the weekly report includes aggregate data on arrests for drug sale or possession with intent to sell. There is no direct data in the weekly report measuring progress on reducing open air-substance use.


3. Lack of shelter and drop-in resources

To help unhoused residents leave the streets, the City must increase the supply of shelter and housing. Unhoused people need a place to go. Figure 5 presents data on the number of weekly exits from street sleeping. The number of people moving to shelter each week fluctuated over the month, starting at 59, dropping to 10, and then increasing each week to end the first week of February at 58. The number of people accessing the Homeward Bound program – which reunites unhoused people with family and friends living outside of San Francisco – has been negligible.



Figure 6, which presents the cumulative number of people exiting from street sleeping, shows that the City has so far helped 180 people leave the streets: 175 moved to shelter, and 5 people accessed the Homeward Bound program.



To evaluate the significance of helping 180 people leave the streets, we need to know how this figure compares to the total number of people who are unsheltered in the Tenderloin. According to the most recent data, from the Point in Time Count for 2019, 1,990 unsheltered people were living in District 6, which includes the Tenderloin. The 2019 report does not specify what share of the unsheltered people in District 6 live in the Tenderloin, but it is likely a majority.


It is unclear how the number of unsheltered people in the Tenderloin has changed since 2019. The pandemic has likely increased the flow of people into unsheltered homelessness, but the City’s use of shelter-in-place hotels early in the pandemic helped many Tenderloin residents leave the streets. It is therefore possible only to conclude that helping 180 people leave the streets is a significant start, but many more people continue to need shelter and housing.


In addition to offering shelter, the City is also providing more drop-in service options. Figure 7 shows the weekly change in meaningful engagements with teams from the Linkage Center, Health Field Outreach, and Homeless Field Outreach. A “meaningful engagement” refers to a conversation in which a City representative provides information about a service but the guest declines to connect to the service. The number of weekly engagements increased over the month of January before decreasing in the first week of February. So far, the City has had 5,631 total engagements.



Figure 8 shows the weekly change in referrals to services from the Linkage Center, Health Field Outreach, and Homeless Field Outreach. A “referral” is a conversation in which a City representative provides information about a service, and the guest states an intention to access the service. Similar to engagements, the number of weekly referrals increased in January before declining in the first week of February. So far, the City has made 2,309 total referrals.



4. Lack of safe passage and accessibility

Figure 9 presents data on two types of infrastructure repairs. The number of weekly repairs to sidewalks and curbs was relatively low, fluctuating between two and four per week. The number of streetlight repairs was similarly low for most of the month before increasing to 72 repairs during the last week of January. DEM has confirmed that the figure is probably an error.



Figure 9 illustrates the typical problem with government data. These two output measures indicate how busy the City has been making repairs. The data does not provide any insight into what we care about, which is whether the City has made progress achieving the priority. After repairing these sidewalks and streetlights, did the City make the Tenderloin’s streets safer and more accessible? How many sidewalks and streetlights need to be repaired? We don’t know.


5. Waste and debris

Figure 10 shows how many tons of waste have been removed from the Tenderloin, and Figure 11 reports the number of power washings. For both sets of data, which come from the Department of Public Works, data is missing for January 16, and the data on waste collected during the week of January 23 is extremely low because it is incomplete.


The data regarding neighborhood cleanliness has the same limitations as the data regarding the fourth priority, above, on establishing safe passage and accessibility. This data presents the City’s outputs – what the City has done to improve cleanliness. However, the data does not directly measure the extent to which neighborhood cleanliness has actually improved.



6. High levels of 911 medical calls

Figure 12 shows that total weekly 911 calls to the Tenderloin declined slightly over the course of the January, from 356 to 280, before increasing slightly to 293 in the first week of February. While the figures are moving in the right direction, more time is needed to see if the trend will continue.



7. Illegal vending

The operating reports on the Tenderloin initiatives do not contain any data specifically about illegal vending.


Conclusion

The City has released weekly operational reports on the Tenderloin Emergency Plan since December, with improvements and modifications made each week. The City’s transparency is commendable and meaningful. Nonetheless, most of the data is limited by the typical problems with government data. The metrics report how busy the City has been, but they do not directly measure the City’s actual progress on achieving its overall objectives.