| Presence of One or More ALPR | |
|---|---|
| (Intercept) | -7.932 |
| (6.694) | |
| log(population) | 0.842 |
| (0.936) | |
| log(popdensity) | -0.346 |
| (0.339) | |
| Business_Percent | -0.049 |
| (0.049) | |
| Residential_Percent | 0.013 |
| (0.017) | |
| log(income) | 0.066 |
| (0.521) | |
| race_nonwhite | 1.243 |
| (1.300) | |
| Freeway_Entrances_Exits | 0.083 |
| (0.120) | |
| log(Total_Crime) | 0.914 *** |
| (0.255) | |
| distance_miles | -0.190 * |
| (0.083) | |
| N | 208 |
| logLik | -87.605 |
| AIC | 195.210 |
| *** p < 0.001; ** p < 0.01; * p < 0.05. | |
Initial Findings - Factors Influencing San Jose ALPR Placement
Introduction
San Jose’s Automatic License Place Reader Camera (ALPR) program has come under considerable scrutiny from local and regional news agencies from its start in 2022. Critics have pointed to possible privacy violations from the warrant-less mass scanning of car license plates, as well as the possible implications of San Jose’s relatively long data retention period (originally 1 year, later shortened to 30 days.) The argument goes that license plate data from these cameras creates a registry of citizens’ travels throughout the city of San Jose, possibly revealing sensitive details of their life, with no due process (ACLU of Northern California 2026.) The AI-enabled features of many of these cameras, allowing simple English language queries to locate cars or individuals, have further fueled this concern. For its part, San Jose’s police force has cited considerable benefits in resolving retail crimes as well as assisting with limited police resources in its justification for placing these cameras (San Jose Police Department 2023.) San Jose’s city government has taken steps to address some of these complaints through reducing the data retention period for ALPR data from 1 year to 30 days, as well as instituting internal audits of police officers access to the data for evidence of non-compliance with California or federal law (ACLU of Northern California 2026.)
Much of the criticism surrounding San Jose’s ALPR deployment relates to their usage, however, substantially less attention has been paid to the locations these cameras were placed. One study, looking at a similar deployment in Hampton Roads, Virginia, found considerable racial and income based disparity in where Flock (a specific brand of ALPRs, also used in San Jose) cameras were placed (Keener et al 2026.) San Jose’s city government has put out very little information as to the reasons for placing cameras in given locations. Seemingly the only publicly available documentation from the city was a press release from the San Jose Police Department in 2023, claiming the cameras were placed to assist with areas of high traffic accidents and organized retail theft. The department also cited limited police resources as a reason for their ALPR deployment (San Jose Police Department 2023.) This study aims to explore the latent factors influencing ALPR camera placement in San Jose, identifying any intentional or unintentional influences on camera deployment locations.
Findings
Table 1 - ALPR Presence versus Demographic/Economic/Criminal Factors
Table 1 shows the results of one of my statistical models, a logistical regression with presence of one or more ALPR camera as the dependent variable. The results are interesting, in that they both confirm and refute San Jose’s official explanation. Perhaps unsurprisingly, logged number of crimes (from 2021 - the year before the first ALPR was deployed) appears to significantly positively correlate with the presence of ALPRs. Each 1% increase in number of crimes is associated with an approximately 2.5 increase in the odds of one or more ALPR being placed in a census tract. Unexpectedly however, distance to the nearest police station negatively correlates with ALPR placement. My reasoning for including such a variable was San Jose City’s claim that ALPRs had been deployed to combat officer resource shortages, which led me to believe that ALPRs would likely be placed in areas further away from existing police stations. But this result seems to possibly contradict San Jose’s public claims. Figure 1 shows this relationship in more detail, showing average number of cameras per census tract at each rounded mile away from the nearest police station rounded. The lack of a significant relationship between the percent of addresses consisting of businesses would also seemingly contradict San Jose’s claim of cameras being deployed to combat retail theft.
Figure 1 - Average Number of ALPRs versus Distance to Nearest Police Station
Also interestingly, no demographic or economic features seem to significantly correlate with ALPR placement. This would seemingly contradict prior research from Keener et al (2026), which found racial and income based discrimination in ALPR placement in Hampton Roads, Virginia. It is possible that differences in policing culture between the two cities or other factors may account for this difference. Number of freeway entrances and exits, a proxy for road traffic, also does not appear to significantly correlate with presence of ALPRs.
These results are still preliminary and I am considering how to best adjust this model to account for factors influencing ALPR placement. Any advice would be very welcome.
References:
ACLU of Northern California. 2026. “SIREN, et al. v. City of San Jose, et al.” ACLU of Northern California. Last updated January 20. https://www.aclunorcal.org/cases/siren-v-city-of-san-jose/.
Keener, Steven, John Finn, and Andrew F. Baird. 2026. “Surveillance Inequality: Race, Poverty, and the Geography of Automated License Plate Reader Deployment.” SocArXiv. https://doi.org/10.31219/osf.io/5ckgv_v1.
San José Police Department. 2023. “SJPD Expands ALPR Program Targeting Organized Retail Theft.” News & Announcements, October 26. https://www.sjpd.org/Home/Components/News/News/1441/262.