Product Sales Analysis

This project is associated with the Master of Science Business Analytics program (Class of 2024) at University of Washington, Seattle.

What I’m most proud of is that I helped my team to navigate unexpected difficulty in processing large amount of data. By developing Python scripts, I helped reduce data processing time from ~1 hour to 5 minutes through an ETL data pipeline.

Olist is a Brazilian e-commerce company that provides a marketplace platform connecting small businesses and consumers. The company has provided real, anonymised commercial data on Kaggle, including 100K orders made from 2016 to 2018. (Data Source: Kaggle)

100K

orders made from 2016 - 2018

$3.2M

in sales revenue

9

data files

Project Goal: understand which product categories are more prone to customers dissatisfaction and how they can improve product sales.

To achieve this goal, we need to:

• Gather and explore data from 9 data sources provided
  
• Extract relevant business insights: identify underperforming product categories and conduct root cause analysis
• Recommend actionable strategies to improve sales
  

To produce business insights on product quality, I have:

• Took initiatives to outline execution strategy and delegate tasks to team members based on strengths and preferences.
• Used Python coding to create an ETL pipeline for processing these 9 data tables. The pipeline was used for:
• Build data schema in MySQL Workbench
  
• Extract and transform data from sources
  
• Load the data to the right destination table
  
• Reduce overall processing time from ~1 hour to roughly 5 minutes and help my team to have data quickly ready in MySQL for analysis purposes. ✨
  
• Used SQL to analyze data for insights after exploratory analysis and brainstorming hypotheses.

Raw Data

Database

Analytics Insights

Recommendation