Flo Customer Segmentation – RFM Analysis

In this project, my goal was to segment customers behaviorally using a large e-commerce dataset and derive business insights from these segments. The workflow was:

1. Data preparation & cleaning: I processed raw sales data, handled missing values, and applied necessary transformations.
2. RFM computation: For each customer, I calculated Recency (time since last purchase), Frequency (purchase count), and Monetary (total spending) values.
3. Segmentation: I used clustering algorithms like K-Means to group customers and compared behavioral patterns of each segment.
4. Valuable segment interpretation: I identified high-value customers and generated insights for marketing strategies.

I visualized analysis results with charts, heatmaps, and generated PDF/HTML reports and dashboard outputs so that business impact was clear.

Project Story and Challenges

When I started this project, I had a large e-commerce sales dataset, but the biggest challenge was that the data was not ready for direct analysis. I performed careful preprocessing to handle data cleaning, missing/outlier values, and correct calculation of RFM parameters. Calculating recency from date formats, normalizing frequency values, and comparing monetary totals across segments were non-trivial engineering tasks.

Setting cluster boundaries and naming segments was also challenging because this stage directly affects the interpretability of insights for business strategy. In the end, the project demonstrated how customer behaviors can be evaluated through segments using concrete data.