al5nfsharyh/wayfair_customer_reviews

---
license: cdla-permissive-2.0
language:
- en
size_categories:
- 10M<n<100M
---

# Wayfair Customer Reviews Dataset

This dataset contains customer reviews collected from **wayfair.com**.
It accompanies the paper *End-to-End Aspect-Guided Review Summarization at Scale*, accepted to the **EMNLP 2025 Industry Track**.

## Overview
The dataset supports tasks such as:

- **Aspect extraction**
- **Product-level summarization**, e.g., aggregating reviews by the `product_id` field.

It can be used on its own or combined with the companion [Wayfair Product Summaries](https://huggingface.co/datasets/IeBoytsov/wayfair_product_summaries) dataset.
Both datasets share a common `product_id` key for easy joining.

## Usage Example

```python
import json
import datasets
import pandas as pd

reviews = datasets.load_dataset("IeBoytsov/wayfair_customer_reviews")["train"].to_pandas()
reviews["aspects"] = reviews["aspects"].map(json.loads)

---
license: cdla-permissive-2.0
language:
- 英语
size_categories:
- 1000万<n<1亿
---

# Wayfair客户评论数据集

该数据集包含从**wayfair.com**收集的客户评论，随论文《大规模端到端方面引导的评论摘要》（*End-to-End Aspect-Guided Review Summarization at Scale*）一同发布，该论文已被**EMNLP 2025产业赛道**接收。

## 概述
该数据集支持以下任务：

- **方面抽取（Aspect extraction）**
- **产品级摘要**，例如通过`product_id`字段聚合评论。

它可单独使用，也可与配套的[Wayfair Product Summaries](https://huggingface.co/datasets/IeBoytsov/wayfair_product_summaries)数据集结合使用。两个数据集共享通用的`product_id`键，便于关联。

## 使用示例

python
import json
import datasets
import pandas as pd

reviews = datasets.load_dataset("IeBoytsov/wayfair_customer_reviews")["train"].to_pandas()
reviews["aspects"] = reviews["aspects"].map(json.loads)