https://github.com/pedramjlo/car_sales_analysis
Car sales analysis
https://github.com/pedramjlo/car_sales_analysis
data-analysis jupyter-notebook pandas python
Last synced: about 1 year ago
JSON representation
Car sales analysis
- Host: GitHub
- URL: https://github.com/pedramjlo/car_sales_analysis
- Owner: pedramjlo
- Created: 2025-02-24T22:09:33.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-03-14T00:24:14.000Z (over 1 year ago)
- Last Synced: 2025-03-14T00:26:45.106Z (over 1 year ago)
- Topics: data-analysis, jupyter-notebook, pandas, python
- Language: Jupyter Notebook
- Homepage: https://www.kaggle.com/datasets/alikalwar/uae-used-car-prices-and-features-10k-listings
- Size: 48.2 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Introduction
This is my personal data analysis project. I downloaded the dataset from Kaggle.
this is not an offical real-world project, but solely for the purpose of self-teaching data analysis and practicing my Python skills.
# Techonologies
- Interactive Python Notebook (Jupyter Notebook)
- Pandas
- Plotly
# The Pipeline
## Data Cleaning (Pandas)
- Imputed values of string (object in Pandas) type containing null values with the mode of the column
- Imputed values of integer/float type containing null values with the mean of the column
- Removed all duplicate rows
- Applied camle-casing on some column header titles
- Capitalised column header titles
- Validated the data type of values column by column
- Normalised Make titles, merged TK with the brands, also abbreviations to full brand names
- Normalised state names and converted them from 2 lower case letter abbreviations to full state names
- Ensured that the Transmission values are either 'Automatic', 'Manual', or 'unknown' (imputed for Nulls)
- Similarly, Ensured that the Color values are either from a list of normal colors or 'unknown' (imputed for Nulls)
- DataCleaner.save_changes() saves all the changes and created a new csv file in './dataset/cleaned/'
## key Analysis Points
- German brands have been doing tremendously. 5/10 selling cars are German:
1- BMW
2- Mercedes-Benz
3- Audi
4- Volkswagen
5- Porsche
- 1990's were bad time in terms of sales, and majority of least profitable cars range from 1991-2004.
- California account for an overwhelming amount of the revenue, $28.81M, followed by Florida, $4.88M, Pennsylvania, $4.81M, and Texas, $2.99M.
- 18 out of 20 dealerships with highest generated are based in California as well as 6 of the least profitable ones.
- R Hollenshead Auto Sales Inc from Pennsylvania and TDAF Remarketing from FLorida are the only non-Californian sellers in the top-20 most profitable sellers.
- A strong positive correlation between number of vehicles and the revenue of a brand. Quantity and consequently more options is a deciding factor.
- Condition of the cars could quite effectively be a strong reason for customers to buy a car.
- The number on odometer is often overlooked by customer and it doesn't influence the sales tremendously.
- US. overseas terittories are amongst least profitable state, including Hawaii, Puerto Rico, and Alaska
- Only 10 states have generated over $1M
- Selling price plays a moderately weak role in the generation of revenue (0.3 coefficiency).
- The Great Depression did not affect our sales significantly. Experiencing a 19.17% incraese in 2007-2008, and only 16.8% decrease in 2008-2009.
- 2010-2011 witnessed a massive spike in revenue increase. Going from $3.06M to $6.6M (53.6%) followed by 26.6% in 2011-2012 ($6.6M to $9M).
- Between 2012 and 2013, sales stalled, but a sharp increase in 2013-14.
- in 2014, a rapid decrease resulted in sales dropping from $12.01M to $3.31M (72.43% decrease).
## Visualisations (Plotly)
