How this was counted

Every number on this site comes from police crash reports. Here is where they come from and what was done to them. The code is public: github.com/mmautner/san-mateo-crashes.

Sources

• Crashes: California Highway Patrol, California Crash Reporting System (CCRS) on data.ca.gov, yearly Crashes, InjuredWitnessPassengers and Parties files, 2017–2025. Downloaded October 15, 2026. Police departments file these reports; CHP publishes them.
• City limits: County of San Mateo GIS, city boundaries layer.
• Council districts: City of San Mateo, council voting districts map.
• Streets and intersections: OpenStreetMap contributors. Map tiles: OpenFreeMap.

Which crashes are included

• Crashes inside San Mateo city limits plus a half-mile band around them, so a circle near the city line also counts crashes just across it in Burlingame, Hillsborough, Belmont or Foster City.
• The counts, map and list cover crashes in which someone was hurt or killed. Each page also says how many crashes police took reports on nearby where no one was hurt (2,709 in 2017–2025, away from freeways). Those have no further detail on the site. Nationally, about 60% of crashes where no one is hurt are never reported to police (NHTSA, 2023), so the real number of these crashes is higher.
• Freeways are left out: crashes on US-101, SR-92 and I-280 and their on- and off-ramps. That removes 5,389 crash reports and 2,382 people hurt (21 killed) from 2017–2025. El Camino Real and other surface streets, including state routes, are kept.
• 2017–2025: 2,384 crashes in which 3,040 people were hurt, 17 of them killed.

The citywide number on the front page counts only crashes inside San Mateo city limits (not the half-mile band), away from freeways.

Who counts as hurt

A person counts when the officer recorded an injury. Witnesses and passengers who were not hurt are not counted.

For deaths, the site shows only the month, year, and whether the person was walking, on a bike or scooter, or in a vehicle.

Crash details

Tapping a crash shows what the officer wrote: the hour, the light, the type of crash, who the officer found most at fault and what they were doing, the cause the officer recorded (translated from the California Vehicle Code section into plain words), and whether it was a hit and run. The Vehicle Code's crosswalk rule covers anyone in a crosswalk, so “didn't yield to someone in a crosswalk” includes a few crashes in which the person hit was on a bike. Fault and cause are the officer's judgment at the scene, not a court finding, and some reports contradict themselves; the site shows them as written. Crashes where someone died show none of these details. The summary of causes on each page counts only the crashes that have details, and a one-line summary appears only when at least 5 crashes are in range and the top cause appears at least 3 times.

Where each crash is placed

Most reports include coordinates. 34 crashes had none and were placed from the report's street description (the two streets, plus the distance and direction the officer recorded). 126 crashes had coordinates more than 200 meters from the streets the officer named; those were moved to the named location. Both kinds are marked “approximate” in the list.

Years

The site starts at 2023 by default. San Mateo Police reports from August 2021 through July 2022 are missing from the state's data, so counts for 2021 and 2022 are too low. Years back to 2017 can be added with the year controls. 2016 is left out because nearly all San Mateo Police reports for that year are missing. The current year is left out until it is complete.

City plans

Each page says whether the city's Local Roadway Safety Plan (adopted May 20, 2024) lists the intersection, or a street segment through it, among its priority locations, and which listed locations fall within the chosen distance. Ranks and names come from the plan's Table 6 (page 48) and Table 7 (page 50). The proposed changes are plain-language summaries of its Table 9 (pages 75–76). The plan ranked locations using crashes from 2017 through 2021, weighted by severity, so its order differs from the counts here. A project's status is shown only when the city reports it complete; otherwise the page links to the city's project page. The source file for these entries is site/plans/lrsp.json in the code repository.

Pages near the Downtown, Hayward Park and Hillsdale Caltrain stations and along El Camino Real also cite the city's Transit-Oriented Development Pedestrian Access Plan (adopted November 21, 2022). Its project list doesn't rank locations. The site reads that list directly from the city's PDF and sorts each location's recommended changes into plain categories by keyword (for example, “LPI” becomes “a head start on the walk signal”), keeping the plan's short-term and medium-term split; each page links to the page of the plan's list with the full wording. Locations that aren't on a street, such as paths across the tracks or station parking lots, aren't placed on pages. The extraction is pipeline/plans_tod.py.

Pages on or near a proposed bikeway also cite the city's Bicycle Master Plan (adopted April 6, 2020), from its Appendix E project list: the street, its limits, the proposed kind of bikeway, and the plan's priority group (High, Medium-High, Medium-Low or Low, from a score for safety, comfort, equity and expected use). The site reads that table directly from the city's PDF. Where a bikeway ends at the city line or a freeway, the end is placed where the street meets it on the map. Projects on paths, overcrossings or parks that aren't streets on the map aren't placed on pages. Some of these bikeways may have been built since 2020. The extraction is pipeline/plans_bmp.py.

Pages also cite the city's 2012 Pedestrian Master Plan where its Appendix G names the location: corners recommended for wider sidewalk corners (Table G-4), stop lines set back from the crosswalk (Table G-6) or a pedestrian hybrid beacon (Table G-7), streets recommended for new sidewalks (Table G-1), and mid-block crosswalk upgrades (Table G-8). These were entered by hand from the plan, with page citations, in site/plans/ped.json. The plan's list of high-visibility crosswalk locations (its Appendix F) isn't included. Since the plan is from 2012, some of these changes may have been made.

Each page sorts the changes these plans recommend at the intersection, or on a plan's street segment through it, by what they take. “No construction needed” is limited to signal timing, signs and paint: for example a head start on the walk signal, no right turn on red, high-visibility crosswalk markings, stop lines set back from the crosswalk, or a painted bike lane. Anything that needs concrete, new signal equipment, lights or beacons is listed as needing construction or new equipment, and so is anything unclear. Studies and enforcement are listed separately. The sorting of every recommendation is in site/plans/kinds.json. The plans don't record what has been done since, so some of these may already be in place.

Distances

Distances are straight-line distances from the point where the two streets meet. Walking times assume about 3 miles an hour: a quarter mile is about a 5-minute walk.

At the intersection

Each page also says how many of the people hurt nearby were hurt at the intersection itself: in crashes placed within 40 meters (about 130 feet) of where the two streets meet. The rest were on the streets around it.

Comparing intersections

An intersection is any place in city limits where two differently named streets meet (1,414 in all). The comparison line counts people hurt within the same distance and years around every one of them. If a location is below the middle, the site shows the typical (median) count instead of a percentage.

What this misses

Only crashes reported to police are here. A UC Berkeley review found that police reports captured 44% to 75% of pedestrian injuries seen at hospitals and 7% to 46% of bicycle injuries, so the real numbers are higher. Nationally, NHTSA estimates that about 32% of crashes in which someone is hurt, and about 60% of crashes in which no one is hurt, are never reported to police (The Economic and Societal Impact of Motor Vehicle Crashes, 2019, page 3). That is why the site says “at least.”