WANDS - Wayfair ANnotation Dataset: Dataset for product search relevance assessment

• Original source of the data is: https://github.com/wayfair/WANDS
• Train, dev, test split of 3:1:1 as per footnote 5 in https://arxiv.org/abs/2307.00370

Details

• 42,994 candidate products
• 480 queries
• 233,448 (query,product) relevance judgements

Column details

• Product columns:
  • product_id - ID of a product
  • product_name - String of product name
  • product_class - Category which product falls under
  • category_hierarchy - Parent categories of product, delimited by /
  • product_description - String description of product
  • product_features - | delimited string of attribute:value pairs which describe the product
  • rating_count - Number of user ratings for product
  • average_rating - Average rating the product received
  • review_count - Number of user reviews for product
• Search queries columns:
  • query_id - unique ID for each query
  • query - query string
  • query_class - category to which the query falls under
• Annotated (product,relevance judgement) pairs, columns:
  • id - Unique ID for each annotation
  • label - Relevance label, one of 'Exact', 'Partial', or 'Irrelevant'

Citation

Please cite this paper if you are building on top of or using this dataset:

@InProceedings{wands,  
  title = {WANDS: Dataset for Product Search Relevance Assessment},  
  author = {Chen, Yan and Liu, Shujian and Liu, Zheng and Sun, Weiyi and Baltrunas, Linas and Schroeder, Benjamin},  
  booktitle = {Proceedings of the 44th European Conference on Information Retrieval},  
  year = {2022},  
  numpages = {12}  
}

Code for generating dataset

import pandas as pd
from datasets import Dataset
from datasets import DatasetDict, Dataset
from datasets import ClassLabel, load_from_disk, load_dataset, concatenate_datasets
from pathlib import Path

base_path = "https://github.com/wayfair/WANDS/raw/main/dataset/"

query_df = pd.read_csv(f"{base_path}/query.csv", sep='\t')
product_df = pd.read_csv(f"{base_path}/product.csv", sep='\t')
label_df = pd.read_csv(f"{base_path}/label.csv", sep='\t')

df_dataset = label_df.merge(
    query_df, on="query_id"
).merge(
    product_df, on="product_id"
)

wands_class_label_feature = ClassLabel(num_classes=3, names=["Irrelevant", "Partial", "Exact"])
dataset = dataset.train_test_split(test_size=2/5, seed=1337)
dev_test_dataset = dataset["test"].train_test_split(test_size=1/2, seed=1337)
dataset = DatasetDict(
    train=dataset["train"],
    dev=dev_test_dataset["train"],
    test=dev_test_dataset["test"],
)
"""
DatasetDict({
    train: Dataset({
        features: ['id', 'query_id', 'product_id', 'label', 'query', 'query_class', 'product_name', 'product_class', 'category hierarchy', 'product_description', 'product_features', 'rating_count', 'average_rating', 'review_count'],
        num_rows: 140068
    })
    dev: Dataset({
        features: ['id', 'query_id', 'product_id', 'label', 'query', 'query_class', 'product_name', 'product_class', 'category hierarchy', 'product_description', 'product_features', 'rating_count', 'average_rating', 'review_count'],
        num_rows: 46690
    })
    test: Dataset({
        features: ['id', 'query_id', 'product_id', 'label', 'query', 'query_class', 'product_name', 'product_class', 'category hierarchy', 'product_description', 'product_features', 'rating_count', 'average_rating', 'review_count'],
        num_rows: 46690
    })
})
"""

dataset.push_to_hub("napsternxg/wands")