Adul Sa-a / Q

Work / Case study

Multi‑Channel E‑commerce Sales Pipeline

FEATURED

หนึ่ง dashboard จาก marketplace exports ที่ไม่เหมือนกัน

Source repository

A reporting view needs comparable platform, order and marketing data without treating incomplete months as completed periods.

My implementation: Power Query/M transformations for orders, ads and CPAS inputs. Documented DAX measure register and dimensional Power BI model.

Repository assets use synthetic/anonymized demonstration data.

E-commerce reporting · Project artifacts
Slide 1 of 3
SCREENSHOT

Committed project artifact. Figures are demonstration data.

Multi‑Channel E‑commerce Sales Pipeline — 01 · Overview KPIs
Power BI e-commerce dashboard showing marketplace filters, reporting KPIs and charts.

Committed project artifact. Figures are demonstration data.

01

Context

Marketplace sales and advertising exports arrive with incompatible columns, status labels, fee fields and reporting periods.

Problem

A reporting view needs comparable platform, order and marketing data without treating incomplete months as completed periods.

What it enables

Creates one comparable reporting layer so a team can inspect channel performance, payout and advertising signals without stitching exports by hand.

Constraints

  • Repository assets use synthetic/anonymized demonstration data.
  • The public repository does not include a Power BI PBIX file, live API integration or scheduled refresh proof.
02

System

Power Query/M pipeline and Power BI model for aligning Shopee, Lazada and advertising exports into one inspectable reporting layer.

DIMENSIONAL ARCHITECTURE

Multi-Channel Conformed Star Schema

Select any table node to inspect its grain, keys, and conformed role.

Fact Table (Measurable Events)

Fact_Orders

Primary Key: order_id

Granular normalized order records across Shopee and Lazada exports.

Field Definitions:
  • order_idPK
  • date_keyFK ➔ JOIN
  • platform_idFK ➔ JOIN
  • product_idFK ➔ JOIN
  • gmv
  • net_payout
  • shipping_fee

Inputs

  • Shopee Seller Center order exports
  • Lazada CSV/Excel order exports
  • Shopee/Lazada ads, affiliate/AMS and Meta CPAS exports

Implementation

  • Power Query standardizes headers, statuses, types, platform and campaign labels.
  • DimDate, DimPlatform and DimProduct connect Fact_Orders and Fact_Marketing in a documented star schema.
  • DAX measures express revenue, net payout, AOV, ad spend and elapsed-day-aware month-over-month comparisons.
  • Power Query/M transformations for orders, ads and CPAS inputs.
  • Documented DAX measure register and dimensional Power BI model.
03

Decisions

  • Compare equal elapsed-day windowsAn incomplete current month should not be compared directly with a completed prior month.
  • Keep raw channel exports separate from semantic reporting measuresSource-specific cleaning stays visible while cross-channel calculations remain consistent in the model.
04

Artifacts

Dashboard overview page with elapsed-day revenue and ad-spend comparison tiles, a daily sales line chart for Shopee and Lazada, a top-five product bar chart and a platform comparison bar.
Multi‑Channel E‑commerce Sales Pipeline: Overview page: equal elapsed-day comparison, daily sales and product mix. Demonstration data.
Dashboard detail page with an hourly order performance combo chart, a platform share donut and a day-of-week by hour order heatmap table.
Multi‑Channel E‑commerce Sales Pipeline: Detail page: hourly and day-of-week order patterns by platform. Demonstration data.
05

Evidence

Evidence classes distinguish what can be inspected directly from what is documented, simulated or reconstructed.

06

Limitations

  • ข้อมูลตัวอย่างเป็น synthetic/anonymized demonstration data — not client performance.
  • No claim of live API ingestion, scheduled refresh, deployment or client outcome.

Discuss a system like this.

Start a work enquiryAll work