Retail Intelligence
Shipping Manifest · Data Warehouse
MNFT-3RMIN
Data Engineering and Analytics

Retail Sales Intelligence Warehouse

9,800 raw transaction rows traced through a real ETL pipeline into a star-schema warehouse — mined for statistical findings and customer segments a business could act on.

View source Interactive BI engine
Cargo Manifest
Total rows9,800
Date range2015–19
Customers793
Products1,861
Avg ship days3.96
Duplicates0
Verified
Item Summary
0
Rows ingested
0 duplicates
0
Unique customers
0
Unique products
0.00
Avg days to ship
Want to run this analysis on your own data?
Upload any retail CSV — KPIs, trends, and RFM segments, entirely in your browser.
Analyze your data
Stage 01 — Received

Ingest

Raw Superstore transaction data — January 2015 to January 2019 — loaded and profiled before any cleaning began.

0
Total rows
0
Columns
0
Missing values
postal code
Column Manifest — 18 fields
Order IDstring
Order Datedate
Ship Datedate
Ship Modestring
Customer IDstring
Customer Namestring
Segmentstring
Countrystring
Citystring
Statestring
Regionstring
Product IDstring
Categorystring
Sub-Categorystring
Product Namestring
Salesfloat
Quantityint
Discountfloat
Stage 02 — Processed

Clean and enrich

Dates standardized, nulls resolved, and four derived fields added. The Days to Ship column turned out to be the project's strongest analytical signal.

FieldIssueResolutionStatus
Postal Code11 missingResolved via City lookupFIXED
Order DateMixed formatsStandardized DD/MM/YYYYFIXED
DuplicatesNot checkedZero exact duplicates foundCLEAN
Ship DateMissing 3 rowsExcluded from ship analysisFLAGGED
Derived fields added
+ Order Year+ Order Month+ Order Quarter+ Days to Ship
Stage 03 — Stored

Star schema warehouse

One fact table, four dimensions. A full 4-way join confirmed every foreign key resolves — 9,800 in, 9,800 out.

0
Fact_Sales
order lines
0
Dim_Customer
0
Dim_Product
0
Dim_Location
0
Dim_Date
Fact_Sales9,800 rowsDim_DateDim_CustomerDim_LocationDim_Product
Integrity check: 9,800 = 9,800 — every fact row resolves across all four dimensions.
Stage 04 — Inspected

Statistical findings

Two one-way ANOVA tests. One confirmed a hunch. One disproved it.

Avg Sales by Region
p = 0.44 — not significant
Days to Ship by Ship Mode
p ≈ 0 — F = 6,950
Shipping speed drives delivery time overwhelmingly — as expected. But it has no relationship with order size (r = −0.006): faster shipping is not preferentially selected for larger orders.
Stage 05 — Delivered

Customer segmentation

RFM metrics, scaled and clustered with KMeans (k=4). Silhouette score of 0.356 — representing genuine customer behavioral separation.

Segment LabelCustomersRatioAvg Spend
High Value637.9%$9,469
Loyal Regular28335.7%$3,298
At Risk34843.9%$1,703
New / Low Engagement9912.5%$1,404
At Riskis the largest segment at 43.9% — representing mid-tier spenders who have gone quiet. Re-engagement should be the highest-priority operational action.