yuvalkorem1/Diamond.Online.Marketplace

Diamond Price Analysis and EDA

Overview

This project explores a real world diamonds dataset with the goal of understanding how physical and qualitative characteristics influence market price.
The analysis includes data cleaning, descriptive statistics, visual explorations, comparisons between diamond types, and extraction of insights.

Main Research Question

How is the price of a diamond influenced by its different characteristics?

Research Sub-Questions

1. How are diamond carat sizes distributed across the dataset?
2. In which price ranges are most diamonds found?
3. Which numeric features have the strongest influence on price?
4. How does price vary between different size categories?
5. What are the price differences between natural diamonds and lab grown diamonds?

Data Cleaning

Several steps were taken:

• Removed three columns with many missing values: Fluorescence, Cut, Culet.
• Removed 3 duplicate rows.
• Removed rows with missing values in essential columns.
• Considered converting categoricals to numbers, but kept the data readable for clear EDA.
• Standardized the Type column (Natural vs Lab grown).

Exploratory Data Analysis

Carat Weight Distribution

Price Distribution by Ranges

Feature Importance

Price Distribution by Size Categories

Natural vs Lab Grown Trend Lines (1 to 2 carat)

Key Insights

• Carat Weight is by far the strongest predictor of price.
• Most diamonds are around one carat and cost between 1000 and 2000 dollars.
• Price variation grows significantly with size.
• Lab grown diamonds are consistently cheaper than natural diamonds.
• The effect of carat weight on price is much weaker for lab grown diamonds compared to natural diamonds.

Uploading and Loading

from datasets import load_dataset
dataset = load_dataset("your-username/diamond-prices")

Files Included

• diamonds (cleaned).csv
• EDA_notebook.ipynb
• README.md