Introduction:

Data was gathered on purchases made at the building-material retailer Home World. All of its customers have membership cards. Moreover, they can become members of the store's loyalty program for $20 per month. The program includes discounts, information on special offers, and gifts.

As an analyst in the team I wanted to study and understand how Home World different outlets performed through the range of dates provided in the data set and give management data driven recommendations.

GitHub Link

Capstone Project - Retail Sales Analysis

Task

As the lead data analyst I was responsible for loading the data, clean the data , perform exploratory data analysis, visualize the data, gather insights and communicate them to stake holders.

Action

I began by creating a Task Decomposition document that would guide me through the tasks and analysis. I then uploaded the data to a Jupyter notebook and using Python and several libraries(pandas, numpy, Seaborn,Plotly) I cleaned the data, performed E.D.A and visualized several graphs .

Loyalty Program Analysis

Result

  • We found that a significant amount of data was missing in the customer_id column, at around 33%, which skewed all the data set toward Shop 0.

  • Later we decided to removed all these rows as it seems that without them all the shops data was uniform and meant that those missing rows were more like a fault in the data collection rather than an anomaly

  • We found that a relative few customers have very big orders

  • By looking for the best selling items we found out that the top ones do not have a price per unit, this leads us to think again about the question of ‘what was the best selling item’ and instead as ‘what are the most profitable items’, this means that the most quantity sold doesn't mean it is the most valuable

  • By looking at the averages of value and orders quantity we found out how effective the Loyalty program is in bringing value not to specifics shops but to the overall franchise

Reflection

I learned that even though we might start with an idea or question that is meant to be answered, through the journey of the analysis more insightful and therefore more relevant questions are raised that lead us to more impactful answers

It is important not to be locked in an idea over where the data might take us, the opposite is true, have an open mind and be open to new questions

When a lot graphs start looking about the same, its time to declutter and rethink what are our goals are.