Contained in this repo are the following:
- A Jupyter Notebook containing any code I've written for this project.
- An "Executive Summary" PowerPoint Presentation that explains the hypothesis tests I ran, my findings, and their relevance to company stakeholders.
- The Northwind SQL Database along with a visual schema of the databases included.
- A py script for Welch's T-Test that I used in the analysis.
- Hypothesis Testin
- Welch's T-Test
- Cohen's d
- ANOVA test
- Tukey's HSD Test
To conduct analysis using statistical analysis, hypothesis testing, and SQL on the Northwind Traders Database to identify areas for improved performance. This included targeted analysis of improvements to their discount system, employee performance monitoring, order processing time per shipping company, and understanding market demand for certain categories of goods.
For this project, I worked with the Northwind database--a free, open-source dataset created by Microsoft containing data from a fictional company. Here's the schema for the Northwind database:
At first, I set out to answer the following question: Does discount amount have a statistically significant effect on the quantity of a product in an order? If so, at what level(s) of discount? I then developed 3 other questions, along with corresponding null and alternative hypotheses per question.
Findings
- With a discount, the average quantity ordered of a product is 27, compared to 22 without a discount.
- US employees manage an average order total of 121.2 orders, compared to UK employees, who average 56 orders.
- Across the 3 Shipping Companies, United Package has the lengthiest processing time of 9.2 days on average, compared to Federal Shipping’s average of 7.5 days.
- There is no interaction between Category and Customer Region regarding Total Revenue per category.