PulsCar analysis of NHTSA public data, September 2026 Topic: the low oil pressure warning -- complaints, recalls and manufacturer communications License: CC BY 4.0. Attribution: "PulsCar analysis of NHTSA public data, September 2026". SOURCE FILES NHTSA Office of Defects Investigation public flat files, downloaded 17 September 2026 into a local SQLite database: - COMPLAINTS: 777,040 owner complaints received 1 January 2015 through 8 September 2026. - RECALLS: one row per make/model/model-year per campaign; campaigns counted DISTINCT. - MANUFACTURER COMMUNICATIONS: one row per make/model/model-year per document; documents counted DISTINCT by NHTSA ID. This is a different and much larger pull than the 74,143-report, ten-nameplate sample the article also cites, which was taken from NHTSA's complaints API on 12 September 2026 for model years 2015-2022. The two are reported separately and are not added together. CASE SENSITIVITY (important) NHTSA complaint narratives are ALL CAPS for filings up to 2020 and mixed case from 2021 on. Every query here is case-insensitive: SQL LIKE (case-insensitive in SQLite) for the first cut, Python re.IGNORECASE for everything after. A case-sensitive uppercase regex, or an uppercase SQLite GLOB, would silently drop most post-2020 complaints -- and in this file the post-2020 years are the majority, because complaint volume on this symptom roughly tripled after 2022. THE COMPLAINT SET 1. SQL: descr LIKE '%oil pressure%' -> 3,557 complaints. 2. Case-insensitive regex requiring the lamp or the condition, not just the phrase: oil pressure (warning |indicator |idiot )?(light|lamp|indicator|warning|message) OR (light|lamp|warning|message).{0,25}oil pressure OR low oil pressure -> 2,107 complaints. 3. Motorcycle and powersports makes removed (KTM, Harley-Davidson, Yamaha, Suzuki, CFMoto, Indian, Triumph) -- 10 rows -> 2,097 complaints. This is the published set. Sample reading: 30 random matches read in full. All 30 concerned an oil pressure warning or a low-pressure condition; no false positives. Three of the 30 described the lamp NOT illuminating, which is counted as its own subset rather than discarded. WHAT THE SET SHOWS - Towed 355 (16.93%), crash 4 (0.19%), fire 28 (1.34%), 2 injuries, 0 deaths, against all-complaint baselines of 5.11% towed, 4.37% crash, 1.82% fire. This is the highest tow share of any complaint set we have published. - The tow share is concentrated: the 529 Ford EcoSport and Focus complaints (recall 23V-905) run 35.35% towed, while the other 1,568 run 10.71%. Both groups describe a seizure or an engine replacement in 27.6% of narratives -- identical. The cluster moves the tow rate; it does not move the severity. - Mileage: 1,152 usable odometer figures; median 78,000; quartiles 53,000 and 106,000; 16.3% at or under 36,000 miles. - 785 (37.4%) describe a stall, shutdown or loss of power. 602 (28.7%) describe a seizure, engine failure or engine replacement. 803 (38.3%) use a tow word in the narrative, against the 16.93% recorded in NHTSA's structured tow field -- the structured field undercounts. - 220 (10.5%) name a sensor, switch or sender. 309 (14.7%) name the oil pump, its belt, its pulley or the pickup. - 201 (9.6%) state that no warning light appeared, or that the oil pressure lamp never came on; 101 of those 201 go on to describe a seizure, engine failure or engine replacement. - Complaints by year received: 74 (2015), 55, 72, 108, 171, 139, 106, 178, 303, 404 (2024), 283 (2025), 204 in the first eight months of 2026. THE CONSUMPTION CROSS-TAB Whole-file query for consumption language (LIKE '%oil consumption%' OR '%burning oil%' OR '%consuming oil%' OR '%losing oil%' OR '%burns oil%' OR '%uses oil%' OR '%quart of oil every%' OR '%excessive oil%') -> 8,338 complaints. Of those, 386 (4.6%) mention "oil pressure" at all and 206 (2.5%) mention the oil pressure lamp. Read honestly, this says what owners wrote, not whether the lamp lit before the failure: an owner describing consumption has no reason to mention a lamp that stayed dark. It is consistent with, and much larger than, the 21-of-385 figure the article cites from its earlier ten-nameplate sample. RECALLS Query: defect LIKE '%oil pressure%' OR consequence LIKE '%oil pressure%', counted DISTINCT by campaign -> 20 campaigns, including two where the WARNING ITSELF was the defect (Volvo 13V-592, Honda 20V-439). MANUFACTURER COMMUNICATIONS Query: summary LIKE '%oil pressure%' -> 635 distinct documents, 2015 to 2026. Theme tagging is keyword-based on the summary text and a document can carry more than one theme: a sensor, switch or sender 230 (36.2%) a DTC quoted 299 (47.1%) software, calibration or reprogram 153 (24.1%) oil pump, pickup or pressure control 136 (21.4%) a leak, cooler or filter housing 111 (17.5%) consumption or oil level 65 (10.2%) bearings, engine damage, replacement 46 (7.2%) LIMITATIONS - Complaints are self-reported and unverified. Counts are NOT failure rates; this file has no denominator of vehicles on the road, and nothing here ranks makes for reliability. - Keyword presence measures what an owner wrote. "Sensor" in 10.5% of complaints does not measure how often a sensor was the fault, and the bulletin file's 36.2% does not either -- bulletins are written for repeatable dealer-visible fixes, complaints are filed when something large happens. The two files have opposite selection biases and are reported that way. - Theme tagging of bulletins is automated from a summary field, not from reading 635 documents. FILES oil-pressure-complaints-by-make.csv makes with 10+ complaints, mileage, tow and severity shares oil-pressure-top-models.csv the 40 most-reported models oil-pressure-context-shares.csv subset counts with the all-complaint baseline oil-pressure-bulletins.csv 635 manufacturer communications with theme tags