San Jose crashes are down. Severe injuries are up 51%.

What the divergence actually means

A 13 percent drop in collisions alongside a 51 percent rise in severe injuries is not a contradiction, and it is not noise. It is the pattern traffic-safety researchers have been describing nationally for a decade: fewer collisions overall, but a larger share of the remaining ones producing life-altering harm.

The usual explanations are vehicle mass and speed. Larger vehicles transfer more energy to whatever they strike, and pedestrians and cyclists absorb that energy directly. Wider, faster arterial roads produce fewer but more severe impacts than dense street grids. Neither factor shows up in a raw crash count, which is why a city can report improving numbers while the injuries getting through the system get worse.

Why the severity-rate table matters more than the raw ranking

King Road and Story Road recorded the most crashes of any intersection in the ranking — 142 — but its severity rate is 6.3 percent. La Ragione Avenue and Tully Road recorded 18 crashes, and 44 percent of them produced a fatal or severe injury. On a raw count, the second location is invisible. On outcome, it is the most dangerous place in the dataset.

That distinction matters for anyone trying to understand their own collision. A crash at a high-volume intersection is a common event with a common outcome. A crash at a location with a severity profile like La Ragione and Tully is not, and the evidence that explains why — approach speed, sightlines, signal timing, roadway design — is the evidence that matters in a claim.

What this means if you were hurt at one of these locations

A documented history of severe collisions at a specific intersection is a fact about the roadway, not about the driver. Where a public entity knew or should have known about a dangerous condition and failed to correct it, California law provides a route to liability under Cal. Gov. Code section 835. That is a demanding standard and it is not established by a crash count alone, but a location with a persistent severe-injury record is where the question is worth asking.

The deadline is the trap. A claim against a city, county or public agency requires an administrative claim within six months under Cal. Gov. Code section 911.2 — not the two years that applies to an ordinary injury claim under Cal. Code Civ. Proc. section 335.1. People routinely discover the roadway-design angle long after the six-month window has closed on it.

Evidence at these locations degrades on the same schedule as everywhere else. Signal timing records, maintenance logs, prior complaint records and camera footage are all obtainable, and all of them are easier to obtain early. The city publishes this crash data weekly; the records that explain a specific collision do not surface on their own.

Use this data

The underlying dataset is public and free. Journalists, neighbourhood associations, safety advocates and city staff are welcome to reuse this analysis with attribution. The figures above can be reproduced directly from the City of San Jose open data portal, and the method section states every choice made in producing them.