Customer Segmentation Using K-Mean Clustering

In this project we are going to distinguish customers based on Recency, Frequency and Monetary. The source code and dataset can access via the provided links below.

This is the data set we are going to investigate the clusters and have 54190 records but need some data cleaning process.

Use Elbow Method to define no of clusters based on our data set and it shows that 3 will be best fit for our clusters. In result, we have got Silhouette Score of 0.74.

Final Conclusion

cluster 0 - customers who shoped very recent (within past 60 days) but not very frequently and spend moderate which is our Normal Customers.

cluster 1 - customers who shoped from long ago and shop freqeuntly and spend moderate which is our Old Customers.

cluster 2 - customers who shoped very recent but not very frequently and spend much overs the shopping which is our Premium Customers and New Customers.

kaggle source code : https://www.kaggle.com/code/kendricklinn/supermarket-customer-segmentation

data set : https://www.kaggle.com/code/fabiendaniel/customer-segmentation/input