Customer Purchase History UK

Columns

• InvoiceNo: A 6-digit number uniquely identifying each transaction. If the code begins with the letter 'c', it indicates a cancelled transaction. This is a nominal field [1]. There are 25,900 unique invoice numbers [2].
• StockCode: A 5-digit number serving as a unique identifier for each distinct product or item. This is a nominal field [1]. There are 4,070 unique product codes [3].
• Description: The name of the product or item sold. This is a nominal field [1]. There are 4,224 unique product descriptions, with some missing values [3].
• Quantity: The number of units of each product included in a transaction. This is a numeric field [1]. Quantities range widely, including negative values, with a mean of 9.55 [4].
• InvoiceDate: The date and time when each transaction was generated. This is a numeric field [1]. The data spans from 1st December 2010 to 9th December 2011 [5].
• UnitPrice: The price per unit of the product in sterling. This is a numeric field [2]. The mean unit price is 4.61 [6].
• CustomerID: A 5-digit number uniquely assigned to each customer. This is a nominal field [2]. Approximately 25% of customer ID values are missing [7].
• Country: The name of the country where the customer resides. This is a nominal field [2]. There are 38 unique countries represented, with the United Kingdom accounting for 91% of entries [7].

Distribution

Usage

• Customer segmentation: Identifying distinct customer groups based on purchasing behaviour using CustomerID.
• Market basket analysis: Discovering product associations and frequently bought together items using InvoiceNo, StockCode, and Quantity.
• Sales trend analysis: Analysing seasonal patterns, peak sales periods, and overall revenue trends over time using InvoiceDate, Quantity, and UnitPrice.
• Geographic sales analysis: Understanding sales performance and customer distribution across different countries.
• Inventory management: Predicting demand for specific products based on historical sales data.
• Churn prediction: Potentially identifying customers at risk of no longer purchasing (though additional features might be needed).
• Fraud detection: Investigating unusual transaction patterns, such as large negative quantities or high-value cancellations.

Coverage

• Geographic: The data primarily covers transactions from the United Kingdom, accounting for 91% of all records. It also includes transactions from 37 other countries [7].
• Time Range: The transactions recorded span from 1st December 2010 to 9th December 2011 [5].
• Demographic: While customer IDs are included, specific demographic details such as age, gender, or income are not part of this dataset. Approximately 25% of customer records do not have an associated CustomerID [7].

License

Who Can Use It

• Data Scientists and Machine Learning Engineers: For developing predictive models related to sales forecasting, customer lifetime value, and recommendation systems.
• Business Intelligence Analysts: For creating dashboards and reports on sales performance, customer trends, and product popularity.
• E-commerce Managers: To inform strategic decisions on pricing, promotions, product bundling, and customer retention campaigns.
• Academic Researchers: For studies in consumer behaviour, retail analytics, and economic modelling.

Dataset Name Suggestions

• UK E-commerce Transactions
• Online Retail Sales Data
• Customer Purchase History UK
• Retail Transaction Log 2010-2011

Attributes