PULSCAR ANALYSIS OF NHTSA PUBLIC DATA — RADIATORS AND INTEGRATED TRANSMISSION COOLERS PulsCar analysis of NHTSA public data, September 2026. License: CC BY 4.0. Files in this folder may be reused with attribution to PulsCar (pulscar.io). -------------------------------------------------------------------------- 1. SOURCE FILES -------------------------------------------------------------------------- NHTSA public flat files downloaded from static.nhtsa.gov and loaded into a local SQLite database: Complaints FLAT_CMPL.txt 777,040 owner complaints (ODINO), received 1 January 2015 - 8 September 2026. The set below runs 1 January 2015 - 8 September 2026. Field definitions: https://static.nhtsa.gov/odi/ffdd/cmpl/CMPL.txt Recalls FLAT_RCL.txt one row per make/model/model year per campaign; campaigns counted DISTINCT on the campaign number. Communications FLAT_TSBS.txt manufacturer communications; counted DISTINCT on the NHTSA document id. Investigations FLAT_INV.txt ODI investigations and defect petitions. Complaints are owner reports. They are not verified repairs, not failure rates, and they carry no denominator of vehicles on the road. A make or model with more complaints may simply have sold more cars, or older ones, or have been the subject of a lawsuit that told owners where to file. Nothing in these files ranks reliability, and nothing here should be read that way. -------------------------------------------------------------------------- 2. THE COMPLAINT QUERY (4,305 -> 3,149 -> 3,134 -> 2,999) -------------------------------------------------------------------------- Step 1 - full-text search (FTS5) of the complaint narrative for the word "radiator" -> 4,305 complaints. Step 2 - failure proximity. At least one occurrence of "radiator" must sit within 100 characters of a failure word: fail | leak | crack | burst | split | rupt | blew | blown | clog | corrod | corros | seam | replac | overheat | broke | broken | hole | seep | pinhole | weep | separat | deterior | rot | defect -> 3,149 complaints. Step 3 - passenger vehicles only. 15 complaints filed under motorcycle and motorhome makes (Suzuki, Harley-Davidson, Ducati, Triumph, Can-Am, Coachmen, Newmar, Thor Motor Coach, Winnebago) were removed. -> 3,134. Step 4 - the fan exclusion, and why it exists. Reading a random sample (below) showed that the largest class of false positives was not noise but a different part: complaints where the only appearance of "radiator" is inside "radiator fan" or "radiator fan shroud", and the part that failed was the cooling fan. We removed the 135 complaints in which EVERY occurrence of "radiator" is immediately followed by fan, cooling fan, shroud or blower. -> 2,999 complaints, filed 1 January 2015 - 8 September 2026. Those 135 are not discarded knowledge: they are the measured size of the confusion the article opens with, in the federal file. A further 115 of the 2,999 kept complaints mention a cooling fan somewhere in the story alongside a genuine radiator failure. -------------------------------------------------------------------------- 3. SAMPLE READING (30 narratives, seed 20260918) -------------------------------------------------------------------------- Thirty complaints were drawn at random from the step-3 set of 3,134 and read in full before any count was published. 26 of 30 genuine radiator failures or radiator-related contamination 3 of 30 the cooling fan or fan shroud was the failed part, and the word "radiator" only qualified the fan -> this is what step 4 removes 1 of 30 the radiator was mentioned as collateral (frame rust that might let the radiator fall out), not as a failure of the radiator False-positive rate before step 4: 4 of 30 (13.3%). Step 4 removes the three fan cases by construction. The remaining class — a radiator named as collateral damage in a larger story — is not separable by query and is left in, so treat every count below as having roughly a 3% ceiling of noise. -------------------------------------------------------------------------- 4. THE CONTAMINATION ("MILKSHAKE") SUBSET — 585 of 2,999 -------------------------------------------------------------------------- A complaint counts as fluid cross-contamination if the narrative matches any of (case-insensitive): transmission fluid .{0,80} (coolant|radiator) (coolant|antifreeze|radiator fluid) .{0,80} (transmission|atf) milkshake | strawberry milk | cross-contaminat atf .{0,40} coolant | coolant .{0,40} atf fluid .{0,30} (mix|mixing|mixed) .{0,40} (transmission|coolant) (transmission|coolant) .{0,40} (mix|mixing|mixed) .{0,30} fluid 585 complaints match. Of those, 525 are Nissan and 512 are 2005-2010 Pathfinder, Xterra or Frontier — the vehicles named in NHTSA defect petition DP12-004. This is a CLUSTER, not a population estimate: a class-action settlement and a widely publicized petition told a specific group of owners where to file, and 56 of the 585 narratives name a lawsuit or class action outright. The correct reading is "here is what the failure mode looks like when it is documented at scale", not "1 in 5 radiator failures contaminate a transmission". Dollar figures inside the contamination narratives: every $ amount between $100 and $60,000 was extracted (250 figures across the subset). Median $4,500, first quartile $3,000, third quartile $5,900; 195 of 250 are $3,000 or more. These are owner-reported amounts in owner-written text — quotes, paid invoices and estimates are not distinguishable from one another. -------------------------------------------------------------------------- 5. MILEAGE -------------------------------------------------------------------------- Reported mileage is used only where it is above 0 and below 400,000. 1,980 of the 2,999 complaints carry a usable mileage. All radiator complaints median 100,000 Q1 67,000 Q3 130,200 Contamination subset (511) median 113,000 Q1 95,000 Q3 138,544 Everything else (1,469) median 93,407 Q1 55,800 Q3 127,249 At or under 36,000 miles 202 of 1,980 This is not a radiator's service life. Complaints are filed when something fails, and a radiator that quietly reaches 150,000 miles never generates a document. The distribution is tilted toward failure by construction. -------------------------------------------------------------------------- 6. MANUFACTURER COMMUNICATIONS -------------------------------------------------------------------------- FTS5 search of the communication summaries for "radiator" returned 601 distinct NHTSA document ids. The same fan exclusion as step 4 was applied to the summaries, leaving 536 bulletins. Of those, 139 (25.9%) describe a leak or coolant loss, 34 name an oil or transmission cooler, and 42 (7.8%) name a warranty coverage, special coverage, service action or rapid service update — that is, a repair someone other than the owner may pay for. -------------------------------------------------------------------------- 7. RECALLS AND INVESTIGATIONS -------------------------------------------------------------------------- 26 distinct recall campaigns name a radiator in the component or defect text, and the largest of them are fan circuits, fan motors, hose clamps and mounting hardware rather than radiator cores. There is no large-population recall of a leaking radiator core in the file. Defect petition DP12-004 (opened 11 June 2012, denied 22 October 2016) is the closest thing to one. It was filed by the North Carolina Consumers Council over MY2005-2010 Nissan Pathfinder, Frontier and Xterra transmission failures caused by the transmission cooler inside the radiator. NHTSA reviewed 2,505 complaints (638 potential hazard, 1,867 customer satisfaction) and denied the petition. -------------------------------------------------------------------------- 8. FILES IN THIS FOLDER -------------------------------------------------------------------------- radiator-complaints-by-context.csv What the 2,999 narratives describe, with counts and shares, plus the NHTSA towed / crash / fire flags against the all-complaint baseline. radiator-complaints-by-make.csv Complaints by make, with the contamination subset broken out. radiator-complaints-top-models.csv The 60 most-reported models, with median reported mileage, how many of those complaints are contamination, and the leading model years. radiator-cooler-contamination-by-model-year.csv The 585 contamination complaints by model year. radiator-manufacturer-bulletins-by-make.csv The 536 radiator bulletins by make, with how many describe a leak and how many name a warranty coverage or service action. No personal data is published: no VIN, no city, no dealer name, no free text. -------------------------------------------------------------------------- 9. LIMITATIONS -------------------------------------------------------------------------- - Complaints are self-reported, unverified, and have no denominator. - Filing is driven by publicity. The Nissan cluster is the clearest example in this dataset of a lawsuit shaping a complaint count. - Keyword matching cannot read intent. A narrative that mentions a radiator while describing an engine failure counts once, the same as a cracked tank. - Mileage and dates are owner-entered and frequently missing. - Bulletin counts are counts of documents, not of affected vehicles.