Wildberries analytics project built in Power BI.
This project demonstrates an end-to-end Power BI analytical implementation:
data modeling → time intelligence → drillthrough navigation → financial validation → data reconstruction under connectivity loss.
This project represents an analytical Power BI report built for a Wildberries (WB) marketplace seller.
The primary goal is to analyze financial performance, identify revenue drivers and losses, and detect inefficient products and logistics operations.
The report is designed to support business decisions in the following areas:
- profitability control
- operational efficiency
- SKU-level performance monitoring
- warehouse and logistics optimization
The report is built on transactional Wildberries sales and realization data for the period:
December 2021 – May 2022
- sales and realization fact tables
- commissions and logistics cost components
- product attributes (SKU, supplier article)
- warehouse and regional dimensions
- calendar dimension with Month-over-Month logic
The semantic layer follows a clean star-schema architecture designed for analytical clarity and stable KPI evaluation.
The model applies controlled filter propagation and avoids ambiguous relationships to maintain deterministic KPI behavior.
- centralized fact tables
- clearly separated dimension tables
- role-playing calendar dimensions for MoM comparison
- explicit relationships to prevent ambiguity in time intelligence
- measure-driven analytical layer (DAX-based KPIs)
The structure supports analytical performance through simplified relationship design and minimized calculation ambiguity.
Time intelligence is implemented using DAX-based measures:
- Previous Month Revenue logic
- Revenue Delta calculation
- MoM percentage growth
- Leader ranking by revenue change
This ensures reproducible period-over-period comparison without ambiguity in temporal context.
Conceptual structure of the semantic layer:
erDiagram
FACT_SALES {
int sale_id
date sale_date
string sku
decimal revenue
decimal quantity
}
FACT_COSTS {
int cost_id
date cost_date
string sku
decimal commission
decimal logistics_cost
decimal cost_of_goods
}
DIM_PRODUCT {
string sku
string supplier_article
string category
string brand
}
DIM_WAREHOUSE {
string warehouse_id
string warehouse_name
string region
}
DIM_CALENDAR {
date date
int year
int month
int month_index
}
FACT_SALES }o--|| DIM_PRODUCT : "by SKU"
FACT_COSTS }o--|| DIM_PRODUCT : "by SKU"
FACT_SALES }o--|| DIM_WAREHOUSE : "by warehouse"
FACT_COSTS }o--|| DIM_WAREHOUSE : "by warehouse"
FACT_SALES }o--|| DIM_CALENDAR : "by date"
FACT_COSTS }o--|| DIM_CALENDAR : "by date"
The model follows a star-schema principle with centralized fact tables and conformed dimensions.
KPI logic is implemented through DAX measures referencing the calendar dimension, ensuring deterministic time-based calculations and stable financial evaluation.
During project execution, access to part of the original data sources was temporarily lost.
Instead of stopping the analytical workflow, the report was reconstructed using:
- cached model artifacts inside PBIX
- preserved transformation logic in Power Query
- recalculated DAX measures
- cross-checking revenue and cost components against prior model outputs
Final financial metrics were reconciled and verified to ensure consistency with previously validated financial aggregates.
All totals were revalidated to eliminate material discrepancies and ensure reconciliation integrity.
This demonstrates architectural transparency and resilience under disrupted data connectivity.
The model includes validation checkpoints to ensure:
- consistency of revenue totals
- alignment of commission and logistics cost components
- reconciliation with validated financial summaries
- integrity of MoM comparisons
Financial KPIs were validated against aggregated totals before analytical conclusions were derived.
All material variances were eliminated after reconciliation.
- revenue
- gross profit
- ROI
- core cost components (WB commission, logistics, cost of goods)
- identification of growth and decline leaders
- comparison with previous periods
- ranking of SKUs by revenue delta
- detailed product-level analysis
- transition from summary KPIs to individual SKU performance
- profitability and ROI by warehouse
- detection of unprofitable logistics directions
- support for warehouse-level optimization decisions
The report is structured for different business roles:
- executives (high-level financial overview)
- analysts (MoM dynamics and comparisons)
- category managers (SKU drillthrough)
Interactive features include:
- synchronized slicers
- cross-filtering between visuals
- drillthrough navigation from summary pages to detailed views
Screenshots of key report pages and their analytical purpose are provided in the docs/screenshots section.
This Power BI report supports structured analytical decision-making for Wildberries marketplace operations.
The dashboard enables:
- rapid assessment of overall business performance
- identification of revenue drivers and loss-making SKUs
- detection of logistics inefficiencies
- warehouse-level profitability analysis
- drilldown from aggregated KPIs to SKU-level detail
It helps answer practical business questions such as:
- Which products caused the largest revenue drop in the selected month?
- Where do logistics costs significantly reduce profitability?
- Which SKUs require pricing or assortment correction?
- Which operational metric most strongly influences profit growth?
Due to GitHub file size limitations, PBIX files are not stored directly in this repository.
Full Power BI report access:
-
Google Drive (PBIX files):
https://drive.google.com/drive/folders/1zoEV2jGBrZNeaEx8QKa3gH0N4XxLiHag -
Power BI Service (private workspace)
- Month-over-Month (MoM) sales analysis
- Leaders of growth and decline
- Drillthrough analysis by SKU
- Clean semantic model and DAX measures
- Financial validation and reconciliation
See key report pages here:
docs/screenshots
-
Finance overview
-
MoM leaders
-
Logistics ROI by warehouse
-
Drillthrough SKU
- Power BI Desktop
- DAX
- Power Query (M)
- Star-schema data modeling
- Semantic layer engineering
pbix/— Power BI report files (metadata; PBIX stored in Google Drive)docs/— business & technical documentationdocs/screenshots/— report screenshots
PBIX files are intended to be opened locally in Power BI Desktop.
This project demonstrates analytical modeling and semantic engineering.
For governance-driven multi-layer BI architecture (TDP → Executive → CFO Snapshot), see:
Marketplace-MultiCab-BI-Case
https://github.com/AvseHorosho/Marketplace-MultiCab-BI-Case
Author: AvseHorosho
GitHub: https://github.com/AvseHorosho