AutoMPG_DataProcessing

Object-Oriented Data Processing Framework for the UCI AutoMPG Dataset

Project Overview

Date: 23 Jan. 2024 - 31 Jan. 2024
Author: Michelle Zong

This project involves creating an object-oriented framework in Python to process and analyze fuel efficiency data from the UCI AutoMPG dataset. The focus of the project is to apply object-oriented programming (OOP) principles to clean, transform, and analyze the dataset efficiently.

Key Features

• Object-Oriented Design: The project is structured around Python classes that encapsulate the data processing logic and methods for efficient data manipulation.
• Data Parsing and Cleaning: Handles non-standard delimiters and missing values in the dataset. Implements flexible data ingestion and transformation methods.
• Iterable Data Structures: Built custom iterable data structures to process the dataset in a memory-efficient manner, allowing for easy exploration and analysis of the dataset.
• Data Comparison: Allows data comparison between two AutoMPG objects such as equality and less than comparisons based on vehicle features (e.g., year, mpg).
• Unit Testing with unittest: The file uses Python's unittest framework to test the functionality of the AutoMPG and AutoMPGData classes.
  • Methods like setUpClass, tearDownClass, setUp, and tearDown are used for setup and cleanup before and after tests.
  • Includes checks for string formatting, equality comparison, less-than ordering, and hashability.

Dataset

• Source: UCI Machine Learning Repository - AutoMPG Dataset
• Description: This dataset contains various characteristics of different cars from the 1970s and 1980s, including attributes such as MPG (miles per gallon), cylinders, horsepower, weight, etc.
• Handling Missing Data: The dataset includes missing values in some columns (e.g., horsepower). These values were cleaned using techniques such as imputation or removal based on analysis requirements.

Project Structure

├── README.md                 # Project documentation\
├── data/                     # Raw and processed datasets\
│   ├── auto-mpg.data.txt     # Original dataset\
│   └── auto-mpg.clean.txt    # Standardized dataset\
├── src/                      # Source code for data processing\
│   ├── autompg.py            # Main script to run the analysis\
│   └── __init__.py\
└── tests/                    # Unit tests for the framework\
    └── test_mpg.py           # Test cases for MPG data processing\

Classes

1. AutoMPG

• Responsibilities: represents make, model, year, and mpg attributes for each record in the data set.
• Methods:
  • __repr__(): Outputs an unambiguous representation of the class object; for developers.
  • __str__(): Human-readable representation of class object; for end-users.
  • __eq__(): Evaluates whether 2 AutoMPG instances are equivalent based on make, model, year, and mpg
  • __lt__(): Evaluates whether an AutoMPG instance is less than the other based on make, model, year, and mpg
  • __hash__(): Makes class hashable again after defining eq method

2. AutoMPGData

• Responsibilities: An iterable class that will read in the data file, clean it, and generate a list of AutoMPG objects.
• Methods:
  • __iter__(): makes class iterable
  • _load_data(): Reads in the cleaned data file, and populates self.data with AutoMPG objects.
  • _clean_data(): Reads in messy data, standardizes data rows and stores the cleaned data in auto_mpg.clean.txt.

How to Run

1. Fork the Repository:

   git fork https://github.com/mmzong/auto-mpg-data-processing.git

2. Install Requirements:

   pip install -r <path/to/requirements.txt>

3. Run the Project:

    python src/autompg.py