---

Used Cars

Analyzing Seller Type, Vehicle Type and Year of Registration

By Data Society [source]

---

About this dataset

> The autos.csv dataset is a comprehensive collection of valuable data about used cars, and provides insight into how the cars are being sold, what price they are being sold for, and all the details about their condition. Each ad contains information such as dateCrawled (the date the ad was first seen), name of the car, seller type (private or dealer), offer type, price, A/B testing information , vehicle type, year of registration (at which year was the car first registered) , gearbox type, power output in PS (horsepower) , model of car , howmany kilometers has it driven so far , monthof registration(when it was first registered)(essentially giving us an idea about its age), fueltype utilized by it( petrol/diesel /electricity/lpg etc.), brand name to which it belongs to  notRepairedDamage - if there is any damage on the vehicle that has not been repaired yet. DateCreated gives us information when this particular advertisement was created in ebay or other place where these cars can be posted. The nrofpictures field will give you an estimate regarding how many images have been included with this ad and postalcode contain info regarding area code where car have been posted.. Lastly lastseen give us time estimation when a crawler last scan this particular post online .All these factors are instrumental in determining a suitable price for used vehicles . Meanwhile regression analysis based on average prices related to years can be done from this dataset .So grab your laptop get ready !!!

More Datasets

> For more datasets, click here.

Featured Notebooks

> - 🚨 Your notebook can be here! 🚨!

How to use the dataset

> This dataset is a great resource to begin exploring the factors that affect used car prices. With features such as dateCrawled, name, seller, offerType, price, abtest among other data points it can be used to uncover how different aspects of a vehicle determine the pricing of second hand cars. > > The first step would be to explore and understand what each of these fields represent and have an idea about their importance when pricing a used car. > One might then proceed by plotting distribution plots for numerical variables such as yearOfRegistration with price or bar graphs for categorical fields like fuelType to observe if there is any correlation with price in these variables. Knowing certain key trends can assist in predicting future market prices more accurately than relying on yearly averages of all car values combined - which might give shapes too broad general trends instead precise predictions when working with this dataset alone. > > In addition understanding how long a listing lasts before being sold would give valuable insight into discover how competitive offers should stay when customers come across relevant listings on say ebay or other trading sites that list used cars; this could achieved by utilizing two columns - lastSeen and dateCrawled - to figure out their average lifespan before they were sold out. It's likely that its higher priced counterparts tend to remain listed longer than cheaper listings which quickly disappear after being seen often enough by members in related markets searching those platforms for new vehicles up for sale at any given time within certain parameters established such as location or age amongst others . > > Finally one might use supervised learning algorithms such as Linear Regression or Random Forest coupled with feature engineering methods like PCA (Principal Component Analysis) aiming at reducing high dimensionality issues on datasets composed mostly of categorical variables so we can perform actual machine learning operations over extracted numerical feature columns from processes along those lines previously mentioned

Research Ideas

> - Analyze the relationship between car prices and age (year of registration) using a linear regression model to suggest which cars provide the best value for money. > - Use classification models to predict vehicle types based on features like powerPS, price, brand, etc. > - Compare and contrast seller types (private vs dealer) by analyzing prices, seller locations and other geographic information in order to give advice on which type of seller provides the best deals for customers

Acknowledgements

> If you use this dataset in your research, please credit the original authors. > Data Source >

License

> > > License: Dataset copyright by authors > - You are free to: > - Share - copy and redistribute the material in any medium or format for any purpose, even commercially. > - Adapt - remix, transform, and build upon the material for any purpose, even commercially. > - You must: > - Give appropriate credit - Provide a link to the license, and indicate if changes were made. > - ShareAlike - You must distribute your contributions under the same license as the original. > - Keep intact - all notices that refer to this license, including copyright notices.

Columns

File: autos.csv

Column name	Description
dateCrawled	Date the car was crawled. (Date)
name	Name of the car. (String)
seller	Type of seller (private or dealer). (String)
offerType	Type of offer (e.g. sale, repair, etc.). (String)
price	Price of the car. (Integer)
abtest	Test type (A or B). (String)
vehicleType	Type of vehicle (e.g. SUV, sedan, etc.). (String)
yearOfRegistration	Year the car was registered. (Integer)
gearbox	Type of gearbox (manual or automatic). (String)
powerPS	Power of the car in PS. (Integer)
model	Model of the car. (String)
kilometer	Kilometers the car has been driven. (Integer)
monthOfRegistration	Month the car was registered. (Integer)
fuelType	Type of fuel (e.g. diesel, petrol, etc.). (String)
brand	Brand of the car. (String)
notRepairedDamage	Whether or not the car has any damage that has not been repaired. (String)
dateCreated	Date the car was created. (Date)
nrOfPictures	Number of pictures of the car. (Integer)
postalCode	Postal code of the car. (Integer)
lastSeen	Date the car was last seen. (Date)

Acknowledgements

> If you use this dataset in your research, please credit the original authors. > If you use this dataset in your research, please credit Data Society.