Personal Data Analysis Project

Global E-Commerce Sales 2025

I explored a dataset of 10,000 global e-commerce transactions spanning 2021 to 2024 across 18 countries. My goal was to analyze revenue trends, customer geography, category dynamics, and pricing behavior to uncover what the data actually tells us.

10,000 Records|18 Countries
Electronics Dominance:Generates ~64% of total revenue ($3.38M) due to high unit prices ($770 avg).
Uniform Purchasing:Average order quantity is constant (~2.7 units) across all price tiers.
Temporal Coverage:Peak volume occurred in Oct 2023; late 2024 reflects dataset sampling cutoff.
Filters:
Showing 10,000 / 10,000 orders
Total Revenue
$5,284,388
Total gross sales value
Total Profit
$1,437,638
Overall Profit Margin27.2%
Total Orders
10,000
Distinct order transactions
Units Sold
27,313
Avg 2.7 units / order
Avg Order Value
$528
Avg Unit Price: $236

Sales & Transaction Volume Over Time

Tracking monthly sales growth and order counts across 48 calendar months.

What I Noticed

Transaction volume grew steadily from early 2021 through late 2023, peaking in October 2023 at $243.9k revenue (427 orders). The decline in late 2024 (down to 12 orders in Dec 2024) is a direct consequence of dataset sampling cutoff, not an organic business collapse.

Geographic Purchasing Patterns

Analyzing sales distribution, order density, and order value across 18 countries.

Regions:
North America
Middle East
Europe
Asia
What I Noticed

Order counts are evenly distributed across regions (North America: ~2,600, Middle East: 2,458, Europe: 2,446, Asia: 2,494). However, Egypt has the highest Average Order Value ($715.80), whereas UAE has the lowest ($437.91) due to differences in customer product basket selection.

Product Category & Subcategory Share

Comparing category gross revenue, order volume, and subcategory sales.

Filter Subcategories:
Electronics
1,977 orders • Avg $770
$3,382,028
64.0%
Home & Kitchen
2,008 orders • Avg $198
$892,842
16.9%
Clothing
1,977 orders • Avg $91
$407,472
7.7%
Books & Media
2,043 orders • Avg $81
$386,644
7.3%
Beauty & Health
1,995 orders • Avg $48
$215,401
4.1%
What I Noticed

Order counts are nearly identical across all 5 main categories (~2,000 orders each). However, Electronics commands 64% of total revenue due to a high average unit price ($770.04), whereas Beauty & Health generates only $215.4k despite receiving 1,995 orders.

Pricing, Quantity & Discount Behavior

Exploring relationships between unit price, volume purchased per order, and discount rates.

Analytical Note on Price Elasticity & Quantity Behavior

The scatter plot and price band breakdown demonstrate that average quantity purchased remains essentially flat at ~2.7 units per transaction across all price levels (corr = -0.002) and discount tiers (corr = 0.0017). Because this cross-sectional dataset lacks controlled temporal price variation or demand shift indicators, it does not support estimating a causal price elasticity of demand.

Order Fulfillment & Return Rate

Cancelled9.1%
$499,132
908 orders
Delivered62.7%
$3,308,170
6,273 orders
Processing9.6%
$537,441
962 orders
Returned18.6%
$939,644
1,857 orders
Return Observation: The dataset records an 18.6% return rate (1,857 orders, ~$980k revenue), representing a significant operational focus point for product quality and customer support.

Customer Segment Performance

Regular
$2,717,416
5,073 orders • Avg $536
New
$1,017,900
1,942 orders • Avg $524
Premium
$974,665
1,961 orders • Avg $497
VIP
$574,407
1,024 orders • Avg $561

Payment Distribution

PayPal
14.7%
Debit Card
14.9%
Credit Card
14.3%
Google Pay
14.2%

Order Transactions Explorer

Browse and inspect detailed order records matching the active dashboard filters.

Showing page 1 of 667 (10,000 records)
Order IDDateCountryCategoryProduct NamePriceQtyDiscountRevenueProfitStatus
ORD-0000012023-01-18MexicoElectronicsXiaomi 13$520.4070%$3,642.80$1,198.69Delivered
ORD-0000022023-11-13USABooks & MediaThinking Fast and Slow$22.59225%$33.88$5.56Delivered
ORD-0000032022-04-28CanadaElectronicsBose QC45$241.3010%$241.30$79.89Processing
ORD-0000042021-10-22JapanClothingSummer Dress$130.9050%$654.50$285.34Processing
ORD-0000052023-08-06FranceClothingAdidas Ultraboost$74.51320%$178.82$60.13Delivered
ORD-0000062021-06-25ChinaClothingPuma Running Shoes$81.89825%$491.34$118.78Delivered
ORD-0000072024-02-07GermanyClothingChino Pants$66.8210%$66.82$19.49Returned
ORD-0000082022-06-05JordanElectronicsGoogle Pixel 7$975.79150%$487.89-$50.80Delivered
ORD-0000092022-11-05FranceBooks & MediaDeep Work$35.25230%$49.35$4.75Delivered
ORD-0000102023-03-14MexicoHome & KitchenBookshelf 5-Tier$110.12140%$66.07-$14.77Delivered
ORD-0000112023-08-14JapanBooks & MediaMicrosoft Office 365$233.6515%$221.97$66.85Delivered
ORD-0000122022-10-23USAElectronicsASUS ZenBook$1,018.4120%$2,036.82$687.73Returned
ORD-0000132023-04-05ItalyClothingAdidas Ultraboost$158.50440%$380.40-$14.50Delivered
ORD-0000142022-06-30USABeauty & HealthVitamin C Serum$14.85215%$25.24$4.74Returned
ORD-0000152023-07-30USAClothingHoodie Sweatshirt$99.84130%$69.89$22.03Delivered
Page 1 of 667

Methodology, Validation & Data Limitations

Data Integrity & Math

I validated all 10,000 transaction records before constructing visualizations:

  • Zero Missing Values: 0 nulls across 26 attributes.
  • Exact Revenue Formula: Revenue = Unit_Price × Quantity × (1 - Discount) matched with 0.00% variance.
  • Profit Verification: Profit = Revenue - Cost verified for all orders.
Price Elasticity Honesty

In cross-sectional analysis, corr(Unit_Price, Quantity) is -0.0021 and quantity per order is constant (~2.7 units):

  • This dataset lacks controlled temporal price variation per SKU.
  • Calculating a naive elasticity formula produces artificial near-zero elasticity.
  • I explicitly treat price-volume behavior as descriptive rather than causal.
Forecasting & Time Cutoff

Monthly order volume peaks in October 2023 (427 orders) and drops off to 12 orders in December 2024:

  • This late 2024 drop represents a dataset sampling cutoff.
  • Extrapolating a predictive sales forecast would introduce heavy truncation bias.
  • Historical trends are presented strictly as observed historical performance.