Among the terrific data sets that NOPD and the City of New Orleans have rolled out in 2016 is one that I haven’t really utilized yet is the Body Worn Camera Metadata database. This database provides information on every recording NOPD has made from a body camera starting in April 2014. The data is a bit messy, but it can help provide an interesting look into how often Calls for Service are being recorded by body cameras.

NOPD’s body worn camera policy can be found here though I’ve taken a screenshot of the relevant portion on when cameras must be turned on below.

To figure out when cameras are being used I downloaded the data and compared the number of calls with at least one recording in 2016 to the overall number of calls (note: I removed incidents with small samples of 15 or fewer calls).

Overall, a body camera appears to make a recording in 64 percent of 2016 incidents through the end of July. That number jumps to 81 percent of 2016 incidents that receive a ‘Report to Follow’ (RTF) disposition suggesting that the most serious incidents are being recorded the vast majority of the time.

A few explanations jump out as to why every incident is not being recorded:

  1. The body cam data set is somewhat messy, so it’s very possible that mistakes in identifying item numbers led me to miscount the exact number for each incident type.
  2. NOPD isn’t the only law enforcement agency responding to Calls for Service and there’s no way to tease out non-NOPD calls from the publicly available data. The State Police answers calls throughout the city as do private security districts. These non-NOPD officers are under no such requirement to use body cameras so that will skew the data at least a bit.
  3. Calls are being taken over the phone which would also reduce the number of incidents in which a recording is being made.
  4. Not all officers wear body cameras. Certain specialty types of incidents are likely handled by officers who do not patrol and are not required to wear body cameras regularly.

Indeed breaking down incidents by whether body cameras are recording confirms that nearly all of the most serious incidents are being recorded. Below is a table with the incident types with the 15 highest percentages of recorded video.

Incident Type Video %
HOMICIDE BY SHOOTING 100.0%
AUTO ACCIDENT FATALITY 100.0%
AGGRAVATED BURGLARY 98.5%
SIMPLE ASSAULT DOMESTIC 98.4%
ARMED ROBBERY WITH KNIFE 96.9%
UNCLASSIFIED DEATH 94.9%
CARJACKING- NO WEAPON 94.1%
CARJACKING 93.0%
SUICIDE 92.9%
DOMESTIC DISTURBANCE 92.6%
SIMPLE BATTERY DOMESTIC 92.3%
ARMED ROBBERY WITH GUN 91.7%
AGGRAVATED ASSAULT DOMESTIC 91.1%
ARMED ROBBERY 90.2%

There are also 11 incident types that a recording is made 100 percent of the time the incident receives an RTF disposition (noise complaints, aggravated battery by shooting, aggravated battery domestic, homicide by shooting, aggravated burglary, suspicious package, officer needs assistance - life in danger, cruelty to animals, auto accident fatality, carjacking - no weapon, and abandoned vehicle). A total of 52 incident types have a recording made in over 90 percent of incidents with an RTF disposition.

The incident types that are being recorded the least also seem to make sense. As seen below these are the types of follow-up incidents that most logically would be handled by an NOPD officer that is not wearing a body camera.

Incident Type Video %
SEX OFFENDER REGISTRATION CHECK 1.6%
MEDICAL SEXUAL ASSAULT KIT PROCESSING 2.0%
WARR STOP WITH RELEASE 7.1%
HOMELESS 11.7%
RETURN FOR ADDITIONAL INFO 15.8%
MUNICIPAL ATTACHMENT 16.2%
TRAFFIC ATTACHMENT 16.8%
FUGITIVE ATTACHMENT 20.4%
ISSUING WORTHLESS CHECKS 21.9%

The real question I have going forward is whether 64 percent of incidents being recorded is good. It makes sense and seems to indicate the policy is being implemented properly, but the only way to know for sure is to follow the issue over time.

For now 64 percent of all calls and over 80 percent of RTF calls can be our baseline.