Ariela Lopez

When the Yale police union’s billboard truck circled Old Campus during first-year move-in this August, one message was adorned with cartoon bullet holes: police ShotSpotter technology had detected over 1,500 gunshots across New Haven in 2024 thus far. 

Most of the detected gunshots are far from campus, but the Yale Police Department is notified of them nonetheless through the city’s complex ShotSpotter system.

ShotSpotter, an audio-based gunfire detection technology first implemented in New Haven in 2009, notifies officers from the New Haven Police Department and the Yale Police Department of suspected shootings. 

The system is run by a company SoundThinking, whose analysts analyze the audio data in real-time from California and Washington, D.C.

“We choose to use [Shotspotter data on our billboards] because we think it’s an important piece of information for the public to know,” union leader Mike Hall told the News. “Like we said, there were 1,500 rounds fired within city limits this year alone. You take into account the possibility of a stray bullet, anything can happen.”

The Yale Police Benevolent Association — the police union — is in its 19th month of tense contract negotiations with the University. It has a long history of branding the city as dangerous during active negotiations. 

There have been no confirmed shots fired in the city’s District 1 — which includes Old Campus and all Yale residential colleges except Pauli Murray and Benjamin Franklin — since the beginning of January 2024, according to the NHPD’s most recently released CompStat report. But Hall believes that such a statistic — though an achievement for the YPD —  assumes that Yale students would be unaffected by crime in the city at large.

The 1,581 ShotSpotter-detected rounds advertised by the YPBA during their first-year move-in action collectively make up around 350 shooting instances.

ShotSpotter technology, used nationwide, operates through a network of audio sensors spread throughout the city. When one of the sensors detects a sound resembling a gunshot, it sends a signal to one of ShotSpotter’s two instant review centers, which are located in Fremont, California and Washington, D.C. There, a human analyst determines whether the notification is plausibly the sound of one or multiple gunshots — a process that takes under a single minute, according to NHPD Sergeant Jarrod Boyce, the department’s liaison to ShotSpotter.

New Haven-based officers then receive a notification on their ShotSpotter mobile apps, which includes the time of the alleged shot, its approximate location and the number of rounds picked up by the system. Officers can also play the audio, which can help them learn more about the nature of the alleged shooting and make determinations about how to respond to the situation.

Boyce told the News that the YPD officers are also notified by ShotSpotter’s system, but that he believes the YPD has its own liaison to SoundThinking, ShotSpotter’s parent company. YPD officers respond to instances of gun violence when they occur on Yale’s campus.

The University’s Office of Public Affairs and Communications confirmed to the News that the YPD has access to ShotSpotter data but declined to respond to further inquiries about how the department acquires or uses it.

Boyce told the News that he finds the system to be accurate in identifying gunshots a majority of the time. He estimated that around 80 to 90 percent of the notifications officers receive on their mobile apps are in fact gunfire, while the remaining “false negatives” could be vehicles backfiring, fireworks or a similar sound.

Thomas Chittum, a spokesperson for SoundThinking, said that ShotSpotter nationally guarantees its clients a 90 percent accuracy.

But when the city was last considering expanding ShotSpotter use in 2022, the NHPD shared that from the 698 ShotSpotter ​notifications that the police department received in 2021, 17 percent were not gunshots. 

Boyce told the News that he notifies his contact at SoundThinking when the system incorrectly indicates a gunshot. 

Since its implementation in 2009, New Haven’s ShotSpotter apparatus has been expanded three times — most recently in 2022 through the addition of sensors. Boyce emphasized that the sensors are spread across the city, covering most of its highly populated areas, and originally implemented in areas where the department had historically seen the highest rates of gun violence.

Boyce believes that the system’s accuracy has increased with each expansion, as an increased number of angles helps pinpoint the source of a gunshot-like sound.

Accuracy aside, Boyce said that ShotSpotter is a reliable reporting system in a city where, he believes, residents might not always report instances of gunfire on their own accord.

“I don’t know if it is some sort of fatigue from there being gunfire in that area, or they’re used to it, or they think it might be something different,” he said of New Haveners declining to notify his department of a shooting. “If it doesn’t affect them, then they’re not particularly inclined to call the police.”

ShotSpotter sensor locations are hidden from the general public, which Chittum said is meant to protect the individuals or organizations that allow ShotSpotter to use their buildings to house sensors.

The company has been criticized in the past for placing its sensors disproportionately in low-income neighborhoods or communities of color in major cities, allowing police departments and their technology a greater opportunity to allegedly surveil marginalized groups.

SoundThinking denies that the locations of sensors are determined by race or income.

“Occasionally people will say, oh, you know, ShotSpotter is only in certain communities, and my response is always, well, that’s probably because that’s where the majority of gun victimization has concentrated,” Chittum said. “I would love to see ShotSpotter spread all the way across every city. But of course, there is a diminishing return on investment if you have a system deployed in a place to detect gunfire where there is no gunfire.”

According to a 2024 Wired investigation that analyzed the leaked locations of ShotSpotter sensors nationwide, 108,131 New Haveners live in a neighborhood with at least one ShotSpotter sensor. Of that number, 24.8 percent are white, 36.2 percent are Black, 30.3 percent are Latine and 4.5 percent are Asian.

Of New Haven’s 2020 population of around 134,000, around 28 percent are white, 30 percent are Black, 31 percent are Latine and 7 percent are Asian — indicating that the sensors are placed in neighborhoods that disproportionately house Black residents, but not by a large margin.

There are 191 ShotSpotter sensor locations in New Haven.

Update, Aug. 30: The article has been updated to include a response from Yale’s Office of Public Affairs and Communications.