Skip to content

TRAP33ZOID/Unsupervised-Learning-Wholesale-Analysis

 
 

Repository files navigation

Unsupervised-Learning-Wholesale-Analysis

This project aims to analyze the purchasing behavior of clients of a wholesale distributor. The dataset contains annual spending in monetary units on diverse product categories. Project Structure:

Data Import:
    Loaded the dataset into Python for analysis.

Data Cleaning:
    Checked for missing or erroneous data.
    Imputed missing values and corrected obvious errors.

Exploratory Data Analysis (EDA):
    Generated summary statistics for each column.
    Visualized data distributions, relationships, and correlations.
    Detected and handled outliers.

KMeans Clustering:
    Preprocessed the data.
    Determined the optimal number of clusters.
    Performed KMeans clustering to segment the data.

Hierarchical Clustering:
    Visualized the hierarchical structure using a dendrogram.
    Extracted clusters from the hierarchical structure.

Principal Component Analysis (PCA):
    Performed PCA to identify the underlying structure of the data.
    Analyzed which combinations of features best describe the customers.

Key Findings:

Customer Segmentation: KMeans clustering identified two main groups of customers with distinct purchasing patterns.
Hierarchical Structure: Hierarchical clustering revealed a clear data structure with potential for multi-level analysis.
PCA Insights: A few principal components explained a significant portion of the variance, indicating feature redundancy.
Correlations: Some product categories, like "Grocery" and "Detergents_Paper", showed strong correlations, indicating co-purchasing trends.

Conclusions:

The analyses provided insights into customer purchasing behaviors, offering valuable information for targeted marketing, inventory management, and other business strategies.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 100.0%