E-Commerce Business Transaction Dashboard

Every online business collects a lot of data without really using it, and that's basically where this project started. The team had access to a year's worth of raw sales data from a UK-based online retailer, somewhere around half a million rows covering everything from individual orders to customer details, product names, and cancellations. It was all sitting in one big spreadsheet with no real structure, so the first job was making sense of it before any actual analysis could happen.

The dates in the file were formatted in a way that could easily be misread, which mattered a lot since getting a month wrong would have thrown off every single trend in the analysis. A handful of customer records were also missing IDs, and cancelled orders weren't marked anywhere obvious either, they were only identifiable through a pattern hidden in the transaction numbers themselves. All of this was fixed properly rather than worked around, since skipping this step usually causes problems later when the numbers don't add up.

Once the data was clean, a full dashboard was built in Power BI to dig into sales trends, product performance, and customer behavior over the year. A few things stood out quickly. Sales spiked sharply heading into November, which made sense given the seasonal timing, but it also meant the business could have planned its stock better ahead of that rush. More notably, two specific products kept showing up both as best sellers and as the most cancelled items at the same time. That pattern isn't something you'd expect to see randomly, so it pointed toward something specific going wrong, most likely stock issues or a quality problem tied to those two items. It was also clear that a small group of customers was responsible for a large share of total revenue, while the vast majority of sales came from just one country, raising an obvious question about where the business might be able to grow if it looked beyond its home market.