Dark Mode
Time Series Sales Forecast Practice
Retail & Consumer Behavior
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This dataset provides synthetic time series data specifically designed for individuals looking to enhance their time series analysis skills. It simulates realistic sales data, incorporating various features such as long-term trends, yearly seasonality patterns, and weekday/weekend effects, along with some noise. The aim is to facilitate practice in predicting the 'number sold' feature. The data models weak correlations between products and stores, offering a practical scenario for forecasting challenges.
Columns
- Date: Represents the specific date in datetime format.
- store: Identifies the unique store.
- product: Identifies the unique product.
- number_sold: Indicates the quantity of an item sold.
Distribution
The dataset is provided as synthetic time series data, split into two main files:
train.csv and test.csv. The test.csv file has a size of 487.83 KB. It is typical for such data to be in CSV format, although not explicitly stated for both files here. The test.csv contains 25.6k valid records across all its columns. Specific row counts for the train.csv file are not detailed in the provided information. The dataset features data for 7 unique stores and 10 unique products, and importantly, contains no null values.Usage
This dataset is ideally suited for anyone wishing to practice and refine their time series forecasting capabilities. It can be used for building predictive models, particularly for sales forecasting. Users can employ metrics such as the Mean Absolute Percentage Error (MAPE) for assessing their solutions and comparing them with others.
Coverage
The dataset spans a 10-year period from 2010 to 2019. The
train.csv file covers the years 2010 to 2018, while the test.csv file exclusively covers data for 2019. The dataset includes information for 7 distinct stores and 10 distinct products. There are no explicit geographic or demographic details provided, as it is a synthetic dataset.License
CC0: Public Domain
Who Can Use It
This dataset is intended for:
- Data Science Learners and Practitioners: To practice and improve their skills in time series analysis and forecasting.
- Students and Researchers: For academic exercises, case studies, and experimenting with various time series models.
- Machine Learning Engineers: For developing and testing predictive models for sales or demand forecasting in a controlled, synthetic environment.
Dataset Name Suggestions
- Synthetic Sales Time Series
- Time Series Sales Forecast Practice
- Retail Sales Prediction Dataset
- Simulated Sales Data for Forecasting
- Sales Activity Time Series
Attributes
Original Data Source: Time Series Sales Forecast Practice
QUALITY