Limitations
There are significant limitations with the source data used in this study. Therefore, the results must be seen with caution and as such, there is no assurance that the results of this preliminary study are appropriate to inform operational decision making or policy.
This study lacks a normative basis against which the source data should be interpreted. For example, I could not find any copy of the metadata catalog issued by the The City of San Jose, nor the SJPD in a manner that conformed with government metadata standards. Therefore, interpretted the data as a ‘best guess’ from the table field names and from the exploratory data analysis.
There is no assurance provided by the City of San Jose in relation to the quality of the data provided. Therefore this study cannot make any assurance on the quality of the source data.
There are patterns of data that were unexpected. At this stage, I do not have enough specific domain knowledge to explain the unexpected pattern. For example,on the week of May 13, 2021, the volume of recorded police calls increase dramatically. One can only guess that policing intensity may have increased due to more operational budget being allocated in that time period. This has influenced my decision to segment time-ordered data from May of every year and to reject the final year's data. 
Address geocoding errors were found due to (a) the use of Common Name in place of Street Names, where Street Names were not provided in the call records. (b) Some Street Names were not standardised. (c) Technical error in the dataset extracted from the Source System (CAD). This means that the study underestimates the number of sites observed.  The level of underestimation has not been quantified.

Conclusions

The large observation sample size provided this study the opportunity to develop robust stochastic models with high confidence. The conclusive results from Borel-Tanner tests in 7 out of 10 trials can safely be interpreted as reasonable evidence in support of the primary hypothesis. The limitations of the source data are likely to be a contributing factor for inconclusive results in the other 3 trials. 
Literature suggests that Borel-Tanner distributions are commonly found in queueing theory and  traffic flow analysis. I speculate that further research into this domain may be able to better determine if call center (and its related call-queuing parameters) operations may be a factor in why such a distribution is observed.