Second American car database

I created “second American car database” in 2017, sourcing data from TheCarConnection.com, and updated until December 2019. Updates are technically still possible, but worthless because the source website is inconsistent and no longer have data for ALL cars.

People interested in new models can buy first American car database which I continue to update every 2 months.

Download free SAMPLES:
American-Car-Database-No-Specs-by-Teoalida-SAMPLE.xls (5 columns)
American-Car-Database-Basic-Specs-by-Teoalida-SAMPLE.xls (28 columns)
American-Car-Database-Full-Specs-by-Teoalida-SAMPLE.xls (210 columns)

Buy complete database + FREE updates for one year:

61 car makes included

Acura, Alfa Romeo, AM General, Aston Martin, Audi, Bentley, BMW, Buick, Cadillac, Chevrolet, Chrysler, Coda, Daewoo, Dodge, Eagle, Ferrari, FIAT, Fisker, Ford, Genesis, Geo, GMC, Honda, HUMMER, Hyundai, INFINITI, Isuzu, Jaguar, Jeep, Kia, Lamborghini, Land Rover, Lexus, Lincoln, Lotus, Maserati, Maybach, Mazda, McLaren, Mercedes-Benz, Mercury, MINI, Mitsubishi, Nissan, Oldsmobile, Panoz, Plymouth, Pontiac, Porsche, Ram, Rolls-Royce, Saab, Saturn, Scion, smart, Subaru, Suzuki, Tesla, Toyota, Volkswagen, Volvo.

History

The first database I made for American market (Year-Make-Model-Trim-Specs offered since 2014) was missing some essential info, for example MSRP (manufacturer’s suggested retail price), in October 2017 one customer interested in MSRP suggested me to extract data from TheCarConnection.com, then in November 2017 another customer asked me for specific technical specs available in TheCarConnection so I included all specifications. After selling privately to 2 customers, both interested in 2017-2018 models only, I decided to scrap ALL data 1990-2018 and make database public for everyone, naming it “Second American car database”.

The URLs were not constant over time, TheCarConnection occasionally deleted cars by merging cars with similar specs. In mid-2019 they added an unique ID for each car, but they deleted nameplates discontinued before 2015 and the December 2019 database had 20% less records than previous one. I provided 2 files for June 2019 (58006 cars) and December 2019 (46107 cars) and some customers had trouble combining new file with older file to have ALL cars in 1 database.

Accuracy notes

Second American database is sold “as it is”, without corrections or additions, because any changes I do, will be lost at next update.

While specifications for each individual car are pretty accurate, filtering cars by class is troublesome due to multiple values with same meaning: “midsize car”, “midsize cars”, “mid-size cars”. Model hierarchy was slightly messed, for example Jetta is 1996-2005, 2009-2014, 2017-2018, New Jetta is 1999, Jetta Sedan 2002-2016, etc (they corrected this in mid-2019).

First American database, have nearly perfect accuracy, constant unique IDs and nothing is deleted. I also improved model hierarchy, corrected errors and expanded database with additional data not available on Edmunds. In May 2018 I made a more complex JSON scraper (instead of HTML) and added MSRP (for 2001-present cars) in first American database, which been outselling second American database 10 times, keeping updates on-going for 2 databases targeting same market became worthless.

The only advantage of second American database is MSRP for 1990-2000 models, gear ratios, dimensions for cargo area / trunk bed (length, width between wheel arches, height to roof), etc. But very few people are interested in these specs.

List of updates

In 2017 TheCarConnection was providing data for 1997-present models. As March 2018 they added 1996 models with specifications and also 1990-1995 models but only naming and MSRP, without specifications. I do not know if in the future, they will add specifications for pre-1996 cars too.

59 makes, 1136 models, 7224 model years, of which 6608 with specs, ? with trims, 46140 trims, as December 2017
61 makes, 1191 models, 8390 model years, of which 8074 with specs, 6644 with trims, 46804 trims, as 12 March 2018, published 18 March.
61 makes, 1210 models, 8651 model years, of which 8313 with specs (? full specs), 56363 trims (? full specs) as 20 September 2018.
61 makes, 1215 models, 8694 model years, of which 8313 with specs (? full specs), 56369 trims (? full specs) as 26 December 2018 (the source website did not added more specs in 3 months?)
61 makes, 1225 models, 8762 model years, of which 8439 with specs (7008 full specs), 57181 trims (49381 full specs) as 1 Feb 2019.
61 makes, 1229 models, 8790 model years, of which 8454 with specs (7023 full specs), 57312 trims (49512 full specs) as 28 March 2019.
61 makes, 1235 models, 8877 model years, of which 8547 with specs (7116 full specs), 58006 trims (50209 full specs) as 4 June 2019.

In mid-2019 TheCarConnection deleted pages of all model nameplates discontinued before 2015. They have 1990 Dodge Grand Caravan is still visible, because a new model with same name is still in production today, but 1990-2007 Dodge Caravan (non-Grand) disappeared because this nameplate is no longer available today. Strange decision!

I attempted to extract all cars and add 2019-2020 models over the 1990-2018 models from June 2019 release but I did not found enough new cars until 18 December when I found more 2020 models, but less 2018 and 2019 models than previous runs of September and December, so beside updating with new cars, they are also doing changes in old cars, deleting / merging 2 similar trims into one. Consequently, I provide 2 separate tables for June and December 2019 updates, until taking further decision.

521 models, 5629 model years, of which 5449 with specs, 44507 trims, as 16 Sep 2019.
525 models, 5696 model years, of which 5460 with specs, 44558 trims, as 07 Dec 2019.
554 models, 5785 model years, of which 5661 with specs, 46108 trims, as 15 Dec 2019.
46 makes, 538 models, 5786 model years, of which 5660 with specs, 46107 trims, as 24 Dec 2019. This was the LAST published update.

TheCarConnection does add new 2021, 2022, 2023 models, but no longer have ALL models. Scraping is technically still possible, and I did ran scraper to get number of models, without scraping each individual trim page to get specifications, because is worthless to publish updates that are missing old models.

38 makes, 472 models, 5510 model years, of which 5392 with specs, 47493 trims, as 30 May 2020. Exotic manufacturers were removed (Aston Martin, Bentley, Ferrari, Lamborghini, Rolls-Royce, etc).
38 makes, 476 models, 5544 model years, of which 5419 with specs, of with 5407 with 47724 trims, as 1 July 2020.
38 makes, 476 models, 5646 model years, of which 5532 with specs, of with 5522 with 49019 trims, as 18 October 2020.
39 makes, 454 models, 3632 model years, of which 3575 with specs, of with 3574 with 33711 trims, as 10 August 2022. Model years older than 2008 were removed.
37 makes, 443 models, 2122 model years, of which 2038 with specs, of with 2034 with 17251 trims, as 30 June 2023.