Project 04: The Iris Complete Walkthrough

Overview

This capstone project applies every technique from every chapter (Ch1–Ch7) to the classic Iris dataset in a single end-to-end analysis. It is the "all chapters in one script" integration project that shows how the techniques connect in a real data-analytics pipeline.

Chapters Combined

Phase	Chapter	Technique
1	Ch 1	Dataset profiling, attribute scale types
2	Ch 2	Univariate descriptive statistics
3	Ch 3	Multivariate statistics, correlation, scatter
4	Ch 4	Preprocessing: missing values, normalisation
5	Ch 5	Clustering: K-means, elbow / SSE
6	Ch 6	Frequent pattern mining: Apriori, association rules
7	Ch 7	Classification: Naive Bayes + k-NN, evaluation metrics

Dataset

Iris — the canonical benchmark dataset in machine learning.

• 150 objects, 4 numeric attributes, 1 class attribute
• Attributes: SepalLength, SepalWidth, PetalLength, PetalWidth (all Ratio scale)
• Class: Species — setosa, versicolor, virginica (Nominal, 50 each)

The Python script uses a hardcoded 30-row representative sample (10 per species) for speed. The R script uses the full built-in iris dataset (150 rows).

Files

File	Description
iris_complete.py	Pure-stdlib Python, 8-section structure, 7 analysis phases
iris_complete.R	Base R + optional arules, 7 matching phases
project_README.md	This file

How to Run

Python

# No external libraries required — pure standard library
python iris_complete.py

R

# Base R is sufficient for most phases.
# Phase 6 (arules) is optional — the script checks and skips if not installed.
# To install arules:  install.packages("arules")
Rscript iris_complete.R
# or inside RStudio: source("iris_complete.R")

Expected Output (Python)

=== PHASE 1 (Ch1): DATASET PROFILING ===
  Attribute      Scale      Discrete/Continuous
  SepalLength    Ratio      Continuous
  ...

=== PHASE 2 (Ch2): UNIVARIATE STATISTICS ===
  SepalLength: min=4.40 max=7.60 mean=5.68 median=5.65 ...
  ...

=== PHASE 3 (Ch3): MULTIVARIATE STATISTICS ===
  Pearson Correlation Matrix (SL, SW, PL, PW):
  ...

=== PHASE 4 (Ch4): PREPROCESSING ===
  Injecting 2 missing values...
  ...

=== PHASE 5 (Ch5): CLUSTERING (K-MEANS) ===
  K=3 cluster assignments: ...
  Purity: ...

=== PHASE 6 (Ch6): FREQUENT PATTERN MINING ===
  Frequent itemsets (min_support=5):
  ...

=== PHASE 7 (Ch7): CLASSIFICATION + EVALUATION ===
  Naive Bayes predictions: ...
  k-NN (k=3) predictions: ...
  Comparison table: ...

Learning Objectives

After completing this project you will be able to:

1. Identify attribute scale types and describe a dataset (Ch1)
2. Compute and interpret univariate statistics (Ch2)
3. Build and read a correlation matrix and ASCII scatter plot (Ch3)
4. Handle missing values and normalise data (Ch4)
5. Run K-means clustering and evaluate with SSE (Ch5)
6. Mine frequent itemsets and extract association rules (Ch6)
7. Train, evaluate, and compare two classifiers (Ch7)