经销商商品销售预测数据集

本数据论文描述了源自厄瓜多尔零售商 Corporación Favorita 的超市日度销售数据集 (2013-2017)，该数据集被用于支持配件供应链的需求预测研究。数据集覆盖 2013 年 1 月 1 日至 2017 年 8 月 31 日，包含 54 家门店约 4000 种商品的 1.25 亿条日度销售记录。数据空间范围为厄瓜多尔 16 个州的 22 个城市，空间精度为门店级。数据内容除主销售数据（销量、是否促销）外，还包含门店信息（位置、类型）、商品信息（类别、是否易腐）、节假日记录、每日交易数及油价等多维辅助特征。数据由 Kaggle 平台公开发布，经过了主办方的清洗与脱敏。本数据集为构建配件供应链中的多模态、上下文感知的需求预测模型提供了丰富的实证基础，适用于供应链研究中的少样本学习、新产品预测和促销效应分析等复杂场景，具有重要的研究与应用价值。

This data paper describes a daily supermarket sales dataset (2013–2017) sourced from Ecuadorian retailer Corporación Favorita, which is used to support demand forecasting research for the accessory supply chain. Spanning the period from January 1, 2013 to August 31, 2017, the dataset encompasses 125 million daily sales records pertaining to approximately 4,000 product items across 54 stores. The dataset’s spatial coverage includes 22 cities across 16 states of Ecuador, with store-level spatial precision. In addition to core sales data (sales volume and promotion indicator), the dataset provides multi-dimensional auxiliary features including store information (location, store type), product information (category, perishability status), holiday records, daily transaction counts, and oil prices. This dataset was publicly released on the Kaggle platform, and has been cleaned and anonymized by the organizing party. This dataset offers a robust empirical foundation for developing multimodal, context-aware demand forecasting models within the accessory supply chain, and is applicable to complex scenarios in supply chain research such as few-shot learning, new product forecasting, and promotional effect analysis, holding substantial research and application value.

搜集汇总

背景与挑战

背景概述

经销商商品销售预测数据集包含厄瓜多尔54家门店约4000种商品的1.25亿条日度销售记录（2013-2017年），涵盖销量、促销、门店信息等多维特征，适用于供应链需求预测研究。