Skip to content

Project 01

Overview

Businesses and organizations often need to understand the relationships between different factors to make better decisions. For example, a company may want to predict the fuel efficiency of a car based on its weight and engine size or estimate home prices based on square footage and location. Regression analysis helps identify and quantify these relationships between numerical features, providing insights that can be used for forecasting and decision-making.

This project demonstrates your ability to apply regression modeling techniques to a real-world dataset. You will: - Load and explore a dataset. - Choose and justify features for predicting a target variable. - Train a regression model and evaluate performance. - Compare multiple regression approaches. - Document your work in a structured Jupyter Notebook.

Dataset

Housing Prices Dataset (Predict home values based on features like square footage and location)
- We use the built-in dataset from scikit-learn:
- from sklearn.datasets import fetch_california_housing
- Additional dataset available on Kaggle:
- Kaggle Housing Prices

Python Library for Machine Learning: scikit-learn

We use scikit-learn, built on NumPy, SciPy, and matplotlib - Read more at https://scikit-learn.org/ - Scikit-learn supports classification, regression, and clustering. - This project applies regression.

Professional Python Setup and Workflow

We follow professional Python practices. Full instructions are available at https://github.com/denisecase/pro-analytics-02/.

Important: VS Code + Pylance may fail to recognize installed packages in Jupyter notebooks.
See the above guides for troubleshooting and solutions.

Project Outline

Machine learning projects follow a structured approach. We will use this approach throughout the course.

Start your notebook professionally with: - a single top-level title - your name (or alias) - the date - a brief introduction that describes the problem and the dataset. - Import the external Python libraries used (e.g., pandas, numpy, matplotlib, seaborn, sklearn, etc.)

Present your work in clearly numbered second-level and third-level headings

Section 1. Load and Explore the Data

  • 1.1 Load the dataset and display the first 10 rows.
  • 1.2 Check for missing values and display summary statistics.

Analysis: What do you notice about the dataset? Are there any data issues?

Section 2. Visualize Feature Distributions

  • 2.1 Create histograms, boxplots, and scatterplots.
  • 2.2 Identify patterns or anomalies in feature distributions.

Analysis: What patterns or anomalies do you see? Do any features stand out?

Section 3. Feature Selection and Justification

  • 3.1 Choose two input features for predicting the target.
  • 3.2 Justify your selection with reasoning.

Analysis: Why did you choose these features? How might they impact predictions?

Section 4. Train a Linear Regression Model

  • 4.1 Define X (features) and y (target).
  • 4.2 Train a Linear Regression model using Scikit-Learn.
  • 4.3 Report R^2, MAE, RMSE.

Analysis: How well did the model perform? Any surprises in the results?

See EXAMPLE_ANALYSIS for more.

README.md (Required)

Include a professional README.md. Include: - a personalized title - an introduction to your project - a clickable link to your notebook file. - Instructions on how to set up your virtual environment and run your notebook locally.

If starting with an assignment README, remove the parts you do not need to present your project.