Embedded Files
10 May 2023
Introduction to the Data
Introduction to the Data
In April of 2023 I started typing all available Call Volume Data (made available through SDFD's ECDC) into a Spreadsheet. As I started to compile the data, I kept a keen eye on the things anyone may want to know: who are the busiest units, who are the slowest, etc. Something even more obvious stood from that spreadsheet. This is the analysis of those numbers.
I assumed that a lot of the analysis that I would end up doing on this site would be subtle and that I would need some tools to help me visualize some of the nuances I'll be looking for. I attempted to make a graph of all of the units for all the months of data I had collected. The graph was going to show the unit Calls per Day over time.
There was a glaringly large dip in call volume I saw amongst some units. Specifically, the units were downtown and it was during the timeframe of August 2019 to July 2020. At that time I was at the Golden Hill Station apart of eastern Downtown and have a distinct memory of why this Call Volume dip occurred. Smart guess if you said COVID, but the real answer is...there was a fire station remodel going on. Fire Station 3 was temporarily closed August 13, 2019 and would not reopen until July 28, 2020. Engine 3 was moved from their home station in Little Italy, and shared Fire Station 4 - closest to Petco Park - with Engine 4.
Calls per Day Data
Calls per Day Data
In the chart below, you'll see the units of Downtown San Diego plotted over time, with their Average Calls per Day represented by the line. The two most significant units to observe are in the most vibrant colors (E3 is in bright magenta and E4 is in bright purple). E7 is a light blue, while the units of Station 1 are labeled with darker blues and the units of Station 11 are labeled with darker greens. You will see that the middle part of the graph stands out. All of the units dip in average call volume (apart from E3) in August 2019 through July 2020. E3 experiences the opposite effect. Those of you viewing this on a desktop will be able to use your mouse to see more detailed data in the chart as well.
I was also curious to see if there was an opposite effect of moving Engine 3. Was there a decrease in average call volume for the units in the area that Engine 3 was being stationed in and what was the increase in call volume like for the units in the area that Engine 3 left behind? Below is a chart like the one above. This chart is for the the three stations that border Fire Station 3. Presumably, these were the stations who became responsible for some of the Call Volume that Engine 3 would have had during the time it was stationed elsewhere. While not as obvious as the dip in Call Volume that we can visualize in the chart above, there is a small rise in Call Volume for these units as seen in this chart.
In the next set of graphs, I have sorted by Calls per Day and group and Unit. Then I divided the graph to show how individual units experienced daily Call Volume before, during and after the geographical move of Engine 3. There are also two different graphs for this type of representation. There is one for the units in the area of Station 4 (where Engine 3 was temporarily relocated to), and one graph for the units around Station 3 (where Engine 3 was temporarily moved out of).
The chart below shows the percentage of Call Volume change that each Unit experienced. For example, it shows that Engine 11 had a decrease in Call Volume of 15%, while Engine 3 was positioned at Station 4. Each unit's "normal" call volume was calculated with the totals for January 2018 through July 2019 and combined with the timeframe from August 2020 through December 2021. This was then compared to the call volume during the timeframe of August 2019 through July 2020 and subsequently expressed as a percentage of change from "normal". A negative percentage illustrates a drop in Call Volume by that percentage, while a positive percentage illustrates a unit experienced and increaes in Call Volume.
Response Time Data
Response Time Data
In union with the "Calls per Day" data visualized above, here are 3 charts that show the Average Response Times in the same fashion as how the data was represented above.
Observations/Considerations
Observations/Considerations
Applying this information to the real world.
Applying this information to the real world.
I think its easiest to analyze these units in terms of the percentage change in their Call Volume. Clearly the call decrease in the part of downtown that gained an extra first responder dramatically decreased. But the reciprocal effect was observed in the part of San Diego that lost a first responder. There are two notable ways of quantifying the effectiveness of the geographical movement of a first responder unit. The first is to estimate the effective change of service provided by the San Diego Fire Department by comparing how response time changed. The second is to estimate how well the perceived workload of the personnel is distributed. These two things will help us illustrate how the people of San Diego can be more effectively served, and then how sustainable that practice is to the workforce.
Summarizing the information.
Summarizing the information.
The final chart displayed in this article shows how Average Response Times decreased by 3.1% as a whole in the downtown area while Engine 3 was located at Station 4. The exception to this were the Response Times for Engine 2, whose Average Response Time increased by more than 5%. Calls per Day for the Units in the downtown area decreased by an average of 2.4, while Calls per Day increased by 0.8 for the observed Units that are outside of downtown.
Other influences that could affect the data.
Other influences that could affect the data.
The fact that this was in the middle of a Global Pandemic raises qualifications to the data. Our triage, treatment and transport differed during the pandemic, with changes to when and why people activated the Emergency Response System. Purely anecdotally (until I can figure out how to confirm this bias with data), "waves" during the pandemic also felt like they affected Call Volume differently. Trends that permeate through the city and isolated time frames don't perfectly capture the makeup of neighborhoods for their population sizes and specific emergency medical needs. By that I mean that population sizes and their EMS demand change but since we don't have access to that data, we cannot control for those changes when observing the Engine 3 relocation.
It is all of these factors that make it difficult to accurately measure how to geographical movement of one unit influences the entire dispatching system in San Diego. With the given data, this was an attempt to explore a case study for the specific effects of this timeframe in our city's downtown neighborhoods.
Other questions raised by this analysis.
Other questions raised by this analysis.
Why did the Response Times for E3 decrease so dramatically? Why have response times for E5 continued to decrease? What (if any) is the correlation between Call Volume and Response Time of a Unit?
Cliff Notes (TLDR)
Cliff Notes (TLDR)
With the addition of a first responder downtown, Daily Call Volume decreased (by 2.4) for all units and Response Times were faster (3.1% faster)
With the loss of a first responder outside of downtown, Daily Call Volume increased (by 0.8) for all downtown stations, but Response Times were still faster (with the exception of E2)
The overall change experienced by the downtown area during these separate time frames shows how the geographical positioning of First Responders can influence the service to the citizens of San Diego and the workload experienced by SDFD personnel.
Page updated
Google Sites
Report abuse