In short: Oakland Unified now has temperature sensors in 80 school buildings, logging readings every 15 minutes. For the first time, the district can see how hot its classrooms actually get. The data shows heat. It does not show what heat does to a nine-year-old doing long division, and it does not cool the room.
At 3:32 in the afternoon on October 6, it was 83°F in Oakland, ten degrees above normal for the date. School was in session.
The outdoor number is the one that makes the news, but it tells you less than you'd think. What mattered inside was what the buildings had been doing with heat for the previous 72 hours.
Oakland Unified has temperature sensors at 80 schools. Every 15 minutes, each sensor logs three numbers: temperature, humidity, and carbon dioxide concentration. The readings are timestamped, and the district calls them public, though pulling room-level data for a specific day turned out to be harder than "public" implies. The dashboard would not produce a downloadable file for this piece. So what follows works from the outdoor temperature record, what's known about how these buildings were constructed, and the physics of heat moving through old masonry, to reconstruct what was probably happening inside on October 6. The room-by-room picture is exactly what the sensors were installed to provide, and future reporting should be able to check this against it.
The building remembers yesterday
Many of these schools went up in the 1950s and '60s: concrete block, brick, thick walls. Those materials take heat in slowly and give it back slowly, a property engineers call thermal mass. A federal guide for school buildings puts the delay at 4 to 12 hours between the outdoor peak and the indoor peak, depending on how thick the walls are and how they were built.
The overnight low going into October 6 was 56°F. That helped, but it wasn't cold enough to reset anything. A teacher unlocking her room at 7:30 a.m. walked into air that had cooled overnight, surrounded by concrete that was still giving off Monday's heat. A thermometer on her desk might have read 70°F. The walls were working from a different number.
The building runs on its own schedule. It collects energy from the sun, through the roof, from thirty kids and a teacher and a projector and a cart of charging laptops, and it lets that energy go on a timeline nobody in the room chose.
A school day's worth of heat
Classrooms don't all heat the same way or at the same rate, and on October 6 the differences would have shown up early.
By mid-morning, east-facing rooms were taking sun at a low angle that slips under the roof overhangs. Federal design guidance notes that east and west windows are harder to shade than south ones for this reason: morning and late-afternoon sun comes in low, beneath overhangs built to block the steep midday angle. OUSD's own project plans prioritize special window film for upper-floor rooms with heavy south and west exposure. The district already knows which rooms cook first.
By early afternoon, with the outdoor temperature past 80°F, the effects stack. Upper-floor rooms soak up heat through the roof all day and pass it straight down. A 2025 study of 3,666 classrooms across 125 Boston public schools (different city, different buildings, same physics) found that on the hottest day, in schools without air conditioning, top-floor rooms averaged 10.3°F hotter than ground-floor rooms. Inside a single building, the spread between the hottest and coolest room was about 25°F. Two second-graders in the same school, one upstairs and one down, are having different days.
One Chabot Elementary teacher described her classroom, small, west-facing, second floor, as a room where heat got trapped. On October 6, that room would have been taking direct afternoon sun right as the outdoor temperature neared its peak, while the roof overhead conducted the day's accumulated heat down into it. Two heat sources arriving at once in a room with no good place to put either.
Then add what the people and equipment produce. Each of thirty bodies puts out roughly the heat of a 75-watt light bulb. There's a projector, and twenty-five laptops charging. With good ventilation that's manageable. With the windows shut because of wildfire smoke or traffic noise, it builds.
Outside, the temperature peaked at 3:32 p.m., more or less when school lets out. The building hadn't peaked yet. Because of thermal mass, the hottest rooms were probably still climbing after the last bus left, unloading stored heat into empty classrooms through the evening. If that night didn't get cool enough to pull the concrete back down, Wednesday morning started from a higher baseline.
When does heat start to matter?
There's no single temperature where learning falls off a cliff — just a handful of findings from studies that measured different things, and they don't stack neatly into a single threshold.
A 2019 review of multiple studies found students did best on school tasks when rooms were below about 72°F. Cooling a room from 86°F to 68°F was associated with roughly a 20% improvement in task performance, but that figure is a modeled average across an 18-degree range, not a switch that flips at a particular number.
Controlled tests with 10- to 12-year-olds found that cooling a room from about 77°F to 68°F helped kids work faster, though not necessarily more accurately.
California's indoor heat illness prevention standard requires employers to act at 82°F, with further protections at 87°F. It covers teachers because teachers are employees. It was written to keep workers from getting heat illness, and it says nothing about whether kids can concentrate.
Roughly, then: learning may start slipping around 72°F, slips measurably more around 77°F, and at 82°F the school has a legal duty to protect the adults in the room. Those numbers come from three kinds of evidence: cognitive performance studies, controlled experiments, and occupational safety law. Lining them up as one scale is tempting, and doing it hides what each of them actually measured and what it can tell you.
Sixteen thousand dollars and a fuse box
The principal at Chabot Elementary told KALW that after one especially hot week, she asked the PTA to buy portable air conditioners for classrooms. Parents raised about $16,000. When teachers ran the units at the same time, the building blew fuses. So they took turns, a few rooms cooling while the rest waited, the way you'd ration anything there isn't enough of.
The mechanics apply well beyond one school.
A portable AC unit pulls a lot of electricity. It plugs into an outlet. That outlet sits on a circuit, a loop of wire that may feed several outlets across more than one room. Every circuit has a breaker or fuse that cuts power if the load goes past what the wire can safely carry; without it, the wire overheats and fires start. All the circuits run back to a main electrical panel, and the building's total electrical service has a fixed capacity set by the panel and the utility connection.
Each link in that chain has a limit, and the smallest one anywhere becomes the limit for the whole system. The parents could buy the machines. The machines needed circuits, the circuits needed capacity, and the capacity wasn't there. OUSD's FAQ acknowledges that many older buildings need major electrical upgrades before modern cooling systems can be installed safely.
A cooling device requires a circuit that requires a panel that requires capacity. Buying any single link doesn't help when the others aren't there.
An earlier Docket piece, "Once a Month, Says the Only Document That Says It," traced the same kind of chain in apartment buildings: a cooling system needs a working unit, wiring rated for the load, a functioning compressor, someone who can pay the electric bill, and someone with the authority to order repairs. Schools have the same structure.
What a cooling room doesn't say
Part of OUSD's response is what the district calls "cooling rooms": supervised spaces, usually a library or office, fitted with window air conditioners, where students and staff can go during extreme heat. As of October, 26 schools had one finished or under construction.
The district's public documents don't say how many students a cooling room holds at once, how long they can stay, what happens to their instruction while they're there, who decides which kids go, or whether the room cools fast enough to help a child who has been sitting in an 87°F classroom since lunch. It's one cool space in a building where most spaces aren't. What that looks like hour by hour, the actual logistics of rotating 400 kids through a room that fits 30, isn't in the public record.
The gap between measurement and relief
Schools were closed October 12 for Indigenous Peoples' Day. An empty building might have shown how much heat the structure produces on its own, with no kids or laptops in it. But October 12 only reached 72°F, eleven degrees cooler than October 6. The weather changed along with the occupancy, and the comparison falls apart. Even with a sensor in every room, a clean experiment needs conditions nobody can control.
That frustration is a small version of the bigger one. The sensors can show which rooms heat first, how fast they heat, and whether a room keeps getting hotter after the outdoor temperature drops. They record a building's thermal behavior in 15-minute increments, which is more than anyone had before.
What they can't capture is everything else happening in the room. Whether there was smoke in the air. Whether a teacher quietly dropped the afternoon math lesson and put on a video because half the class had gone glassy. Whether the kid who left the room went to the nurse or just stood in the hallway, where it was two degrees cooler. What it was like to be nine years old and trying to concentrate, with the second-floor west windows full of afternoon sun and the concrete still letting go of yesterday.
The sensors made the heat visible, and that matters, because a number is harder to argue with than a complaint. But seeing the heat is the first of three problems. The second is knowing what it does to the kids in the room, and the third is cooling the room. The 15-minute readings take care of the first. The other two are where they were before the sensors went in.
-
Better envelopes, longer safety: A 2026 study modeled how much longer homes in six U.S. climate zones stay safe during power outages when walls, windows, and insulation are upgraded — Portland's modeled safe window jumped from 1.1 to 6.8 days, raising the same question Oakland's schools face about what the building itself can do before mechanical cooling arrives.
-
Boston's classroom sensor study: Researchers who monitored 3,666 classrooms across 125 Boston public schools found temperature differences of up to 25°F within a single building, with top-floor rooms in uncooled schools averaging more than 10°F hotter than ground-floor rooms on the worst days.
-
Warnings that erode confidence: A preregistered study tracking 513 at-risk adults during a February 2026 atmospheric river found that receiving more hazard communication increased protective behavior through greater risk perception but simultaneously lowered people's confidence that available actions would actually help — a dynamic worth watching as Oakland's dashboard makes classroom heat newly visible to parents.
-
Window film's mixed reception: KALW reported that teachers at several Oakland schools said the district's window film installations provided little noticeable relief in many rooms, a claim the new sensor data could eventually confirm or complicate with before-and-after temperature traces.
