Dark Mode
Transactional Retail Behaviour Data
Retail & Consumer Behavior
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This transactional data spans two years from a UK-based non-store online retailer specialising in unique all-occasion gift-ware. The company has a significant customer base, including wholesalers. The data includes multivariate, sequential, time-series, and textual characteristics, encompassing details such as invoice numbers, product codes, quantities sold, and customer residency. Analysing this data can assist users in exploring temporal dependencies and predicting retail transaction dynamics over time.
Columns
- InvoiceNo: A nominal, 6-digit integral number unique to each transaction. If the number starts with 'c', it signifies a cancellation.
- StockCode: A nominal, 5-digit integral number that serves as the unique product code for each item.
- Description: The nominal name of the product or item. Note that this feature contains a small number of missing values (4,382 records).
- Quantity: The numeric quantities of each product per transaction. This feature includes negative values, which typically represent returns.
- InvoiceDate: A numeric timestamp indicating the date and time when the transaction was generated.
- UnitPrice: The numeric price of the product per unit, recorded in sterling (£).
- CustomerID: A nominal, 5-digit integral number uniquely assigned to each customer. It is important to note that this identifier is missing for approximately 23% of the records (243,000 records).
- Country: The nominal name of the country where the customer resides.
Distribution
The dataset is structured as tabular data, commonly distributed as a CSV file (e.g.,
online_retail_II.csv). The file size is 94.85 MB. It records 1,067,371 instances (transactions) across 8 distinct features. The data characteristics include multivariate, sequential, time-series, and text.Usage
This data is ideal for several analytical tasks, including:
- Developing machine learning models for tasks such as classification, regression, and clustering.
- Uncovering hidden patterns in customer behaviour and identifying product sales trends.
- Optimising pricing strategies and predicting future sales volumes.
- Segmenting customers effectively for highly targeted marketing campaigns.
- Detecting anomalies or potential fraudulent activities within the transaction logs.
- Studying modern online retail operations.
Coverage
The dataset contains all transactions recorded by the UK-based online retailer between 01/12/2009 and 09/12/2011. Geographically, the transactions originated from 43 unique countries, although the United Kingdom accounts for 92% of all records. Many of the company's recorded customers are wholesalers.
License
Attribution 4.0 International (CC BY 4.0)
Who Can Use It
- Data Scientists: For developing predictive models, such as sales forecasting or customer lifetime value calculations.
- E-commerce Managers: For gaining strategic insights into product performance, seasonal trends, and inventory management decision-making.
- Academic Researchers: For studying modern retail operations, market dynamics, and temporal dependencies in consumer activity.
Dataset Name Suggestions
- Online Retail II Transactions
- UK E-commerce Giftware Sales Log
- Transactional Retail Behaviour Data
- Online Retail II
Attributes
Original Data Source: Transactional Retail Behaviour Data
QUALITYLoading...