Introduction
Virtual streets audits are techniques that enable urban planners, researchers and professional social workers make informed decisions about local communities by the use of digital technologies such as “Google Street View” \cite{Badland2010}. The purpose of such audits is to better support better wellbeing outcomes for urban communities. I suggest that virtual street auditors will be better served with more context-rich information to support their assessments. The objective of this paper is to present the research done to (a) examine the suitability of government open data to provide this additional context. (b) study the quantitative model relationship between physical urban areas and the volume of service delivery provided to those location areas. (c) Visualise the spatial characteristics and trends in context to the services being studied. (d) to conduct analysis consistent with reasonable public expectations of the privacy of individuals and groups of individuals.
Specifically, I examine the model relationship
\(P\ \to S\)p
between P, the location and the Sp, quantity of service delivered to that location.
I formulate the primary hypothesis (H1) that there exists a stochastic quantitative model that reliably characterises the relationship between P and S. My assumption that each SJPD visit is regarded as a discrete and independent event. Therefore, I limit my research to the family of probabilistic discrete distributions. I suggest that a ChiSquare test result with P \(\ge\) 0.05 would be a reasonable confirmation of my primary hypothesis.
In this paper, I examine over 6 million records of publicly available government open data: the San Jose Police Department’s call center data, sourced from the San Jose City \cite{portal} , over a ten year period, starting May 2011 and ending May 2022. The dataset is updated daily, as a time series, with a latency of 1 day.
I define the quantity of service provided on the site as the number of visits that SJPD physically makes to the location address recorded in the call record. My rationale for this is that, this way, there is some reasonable assurance that that the final determination of the outcome of the visit (recorded as a 'final disposition') will be governed by the Department's internal policy.
I define the class of service - "community support service" - as the service requested by the public to support them in the context of "Family Disturbance" and those that are not assessed by the SJPD as resulting in a decision to Arrest, Cite or perform any other serious criminal justice intervention. In other words, I assume that in cases where the SJPD has decided not to take adverse legal action, it is essentially acting in a community support role in predominantly family related matters. I take the disposition of "R - Report Taken" as the threshold at which to include the visit for study. To be clear, I include disposition statuses of "O- Unfounded Event", "G - Gone on Arrival" in the study because these events suggest that they are not trivial and that these disposition codes most probably reflect SJPD administration policy - not necessarily the family's context in question.