Work / Simulated case study

Shopee Thailand Analytics

FEATURED

Marketplace analytics for sales, customers, campaigns and delivery

Source repository

Operational marketplace records need a legible model for connecting sales trends, customer retention, campaign activity and courier performance.

My implementation: MySQL-style SQL analysis Documented Power BI/DAX semantic-model logic

The README describes a simulated 300,000-order case; this is not Shopee internal/client data.

01

Context

A simulated Shopee Thailand marketplace dataset is used as a portfolio case to explore executive-level commercial and operational questions.

Problem

Operational marketplace records need a legible model for connecting sales trends, customer retention, campaign activity and courier performance.

What it enables

Turns separate commercial and logistics questions into a shared analytical model for clearer review of sales, retention, campaigns and delivery.

Constraints

  • The README describes a simulated 300,000-order case; this is not Shopee internal/client data.
  • The public repository has no committed PBIX, automated tests, CI workflow or licence file.
02

System

SQL analysis modules and a documented Power BI/DAX model for examining sales, customer behavior, campaigns and delivery performance.

DIMENSIONAL ARCHITECTURE

Marketplace Analytical Star Schema

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

Fact Table (Measurable Events)

Fact_Orders

Primary Key: order_id

Central transaction grain across the simulated Shopee Thailand marketplace dataset.

  • order_idPK
  • date_keyFK ➔ JOIN
  • customer_idFK ➔ JOIN
  • product_idFK ➔ JOIN
  • campaign_idFK ➔ JOIN
  • gross_revenue
  • order_status

Inputs

  • Simulated orders
  • customers
  • products
  • campaigns
  • shipments
  • reviews

Implementation

  • Documented star schema centered on marketplace orders.
  • SQL modules for sales performance, customer segmentation, campaign effectiveness and logistics performance.
  • README-documented DAX calculations for segmentation, YoY and repeat rate.
  • MySQL-style SQL analysis
  • Documented Power BI/DAX semantic-model logic
03

Decisions

  • Separate commercial and logistics questionsRevenue, retention, campaign and courier questions require different operational views of the same modeled data.
04

Artifacts

Customer analytics page with unique-customer, repeat-rate and revenue-per-customer tiles, a buyer-segment donut chart, a Thailand customer map and total revenue by segment.
Shopee Thailand Analytics: Customer segmentation and retention view. Simulated case-study data, not client results.
Power BI data model screenshot for the simulated Shopee analytics case study.
Shopee Thailand Analytics: Committed data-model artifact.
05

Evidence

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

06

Limitations

  • กรณีศึกษานี้ใช้ข้อมูลจำลอง; displayed figures are not Shopee or client internal results.
  • README-reported figures are not presented as independently reproduced outcomes.

Discuss a system like this.

Start a work enquiryAll work