- Industry: Retail
- GitHub URL: https://github.com/PoornimaKanasen/KPMG_Virtual_Internship
- Tableau URL: https://public.tableau.com/app/profile/poornima.kanasen/viz/KPMG_WORKINPROGRESS-3/RFM-CustomerSegmentation
Objective
To understand the data distributions, feature engineering, data transformations, modelling, results interpretation and reporting.
Completed tasks
- • Data Quality Assessment
- • Data Exploration, Model Development and Interpretation
- • Visualizations and Presentation
RFM Analysis
RFM uses sales data to segment a pool of customers based on their purchasing behavior. The resulting customer segments are neatly ordered from most valuable to least valuable. This makes it straightforward to identify best customers.
Recency stands for how long it takes for the customer to come back to you to purchase. Naturally, the longer it takes, the more chance it is for the customer to drop out or lose interest.
Frequencyis how many transactions are there between the customer and you. Higher frequency means your products are of interest to them and they are valuable customers to the company.
Monetaryhelps us to understand how much a customer spends on average or in total which is the final measure of his or her value.
Tools used
Python, Pandas, Numpy, Matplotlib, Plotly, Seaborn, Tableau
Interpretations
- • The most important aspects of RFM is Recency because you want the customers to keep coming back and purchase products from your company.
- • The customers can be easily segmented based on RFM score. RFM analysis will help us to identify the kind of approach that should be taken for each segment of customers.
- • For the lost customers, more discounts can be allocated while for the loyal ones, point system could be implemented. The more they buy, the more points can be accumulated which at later stage can be used to redeemed in other purchases.
- • The brand/product which is being sold the most can be continued to be promoted to the group of customers who bought them, perhaps, more variety of accessories or upgrades relevant to the brand.
- • Most bikes are used for standard usage and is identified bought the most by female customers. Awareness can be created pertaining to health and benefits of cycling can be further spread. Besides endorsing or giving importance to the health factor, mini competitions can be organized by the company to further promote the brands/company.
Visuals from Python-Jupyter Notebook-Pandas-Matplotlib-Seaborn-Plotly
![](assets/projects/p1_detail_1.png)
![](assets/projects/p1_detail_2.png)
![](assets/projects/p1_detail_3.png)
![](assets/projects/p1_detail_4.png)
![](assets/projects/p1_detail_5.png)
![](assets/projects/p1_detail_6.png)
![](assets/projects/p1_detail_7.png)
![](assets/projects/p1_detail_9.png)
![](assets/projects/p1_detail_8.png)