--- language: - en - es pretty_name: " 💾🏋️💾 DataBench 💾🏋️💾" tags: - table-question-answering - table - qa license: mit task_categories: - table-question-answering - question-answering --- # 💾🏋️💾 DataBench 💾🏋️💾 This repository contains the original 65 datasets used for the paper [Question Answering over Tabular Data with DataBench: A Large-Scale Empirical Evaluation of LLMs](https://huggingface.co/datasets/cardiffnlp/databench/resolve/main/Databench-LREC-Coling-2024.pdf) which appeared in LREC-COLING 2024. Large Language Models (LLMs) are showing emerging abilities, and one of the latest recognized ones is tabular reasoning in question answering on tabular data. Although there are some available datasets to assess question answering systems on tabular data, they are not large and diverse enough to evaluate this new ability of LLMs. To this end, we provide a corpus of 65 real world datasets, with 3,269,975 and 1615 columns in total, and 1300 questions to evaluate your models for the task of QA over Tabular Data. ## 📚 Datasets By clicking on each name in the table below, you will be able to explore each dataset. | | Name | Rows | Cols | Domain | Source (Reference) | |---:|:-------------------------------|-------:|-------:|:---------------------------|:-----------------------------------------------------------------------------------------------------------------------------------| | 1 | [Forbes](https://public.graphext.com/0b211530c7e213d3/index.html?section=data) | 2668 | 17 | Business | [Forbes](https://www.forbes.com/billionaires/)| | 2 | [Titanic](https://public.graphext.com/8577225c5ffd88fd/index.html) | 887 | 8 | Travel and Locations | [Kaggle](https://www.kaggle.com/competitions/titanic/data)| | 3 | [Love](https://public.graphext.com/be7a566b0c485916/index.html) | 373 | 35 | Social Networks and Surveys | [Graphext](https://public.graphext.com/1de78f6820cfd5ba/index.html) | | 4 | [Taxi](https://public.graphext.com/bcee13c23070f333/index.html) | 100000 | 20 | Travel and Locations | [Kaggle](https://www.kaggle.com/competitions/nyc-taxi-trip-duration/overview) | | 5 | [NYC Calls](https://public.graphext.com/1ce2f5fae408621e/index.html) | 100000 | 46 | Business | [City of New York](https://data.cityofnewyork.us/Social-Services/NYC-311-Data/jrb2-thup) | | 6 | [London Airbnbs](https://public.graphext.com/6bbf4bbd3ff279c0/index.html) | 75241 | 74 | Travel and Locations | [Kaggle](https://www.kaggle.com/datasets/labdmitriy/airbnb) | | 7 | [Fifa](https://public.graphext.com/37bca51494c10a79/index.html) | 14620 | 59 | Sports and Entertainment | [Kaggle](https://www.kaggle.com/datasets/stefanoleone992/fifa-21-complete-player-dataset) | | 8 | [Tornados](https://public.graphext.com/4be9872e031199c3/index.html) | 67558 | 14 | Health | [Kaggle](https://www.kaggle.com/datasets/danbraswell/us-tornado-dataset-1950-2021) | | 9 | [Central Park](https://public.graphext.com/7b3d3a4d7bf1e9b5/index.html) | 56245 | 6 | Travel and Locations | [Kaggle](https://www.kaggle.com/datasets/danbraswell/new-york-city-weather-18692022) | | 10 | [ECommerce Reviews](https://public.graphext.com/a5b8911b215958ad/index.html) | 23486 | 10 | Business | [Kaggle](https://www.kaggle.com/datasets/nicapotato/womens-ecommerce-clothing-reviews) | | 11 | [SF Police](https://public.graphext.com/ab815ab14f88115c/index.html) | 713107 | 35 | Social Networks and Surveys | [US Gov](https://catalog.data.gov/dataset/police-department-incident-reports-2018-to-present) | | 12 | [Heart Failure](https://public.graphext.com/245cec64075f5542/index.html) | 918 | 12 | Health | [Kaggle](https://www.kaggle.com/datasets/fedesoriano/heart-failure-prediction) | | 13 | [Roller Coasters](https://public.graphext.com/1e550e6c24fc1930/index.html) | 1087 | 56 | Sports and Entertainment | [Kaggle](https://www.kaggle.com/datasets/robikscube/rollercoaster-database) | | 14 | [Madrid Airbnbs](https://public.graphext.com/77265ea3a63e650f/index.html) | 20776 | 75 | Travel and Locations | [Inside Airbnb](http://data.insideairbnb.com/spain/comunidad-de-madrid/madrid/2023-09-07/data/listings.csv.gz) | | 15 | [Food Names](https://public.graphext.com/5aad4c5d6ef140b3/index.html) | 906 | 4 | Business | [Data World](https://data.world/alexandra/generic-food-database) | | 16 | [Holiday Package Sales](https://public.graphext.com/fbc34d3f24282e46/index.html) | 4888 | 20 | Travel and Locations | [Kaggle](https://www.kaggle.com/datasets/susant4learning/holiday-package-purchase-prediction) | | 17 | [Hacker News](https://public.graphext.com/f20501a9d616b5a5/index.html) | 9429 | 20 | Social Networks and Surveys | [Kaggle](https://www.kaggle.com/datasets/hacker-news/hacker-news) | | 18 | [Staff Satisfaction](https://public.graphext.com/6822ac1ce6307fec/index.html) | 14999 | 11 | Business | [Kaggle](https://www.kaggle.com/datasets/mohamedharris/employee-satisfaction-index-dataset) | | 19 | [Aircraft Accidents](https://public.graphext.com/1802117b1b14f5c5/index.html) | 23519 | 23 | Health | [Kaggle](https://www.kaggle.com/datasets/ramjasmaurya/aviation-accidents-history1919-april-2022) | | 20 | [Real Estate Madrid](https://public.graphext.com/5f83ec219a7ea84f/index.html) | 26026 | 59 | Business | [Idealista](https://public.graphext.com/5f83ec219a7ea84f/index.html) | | 21 | [Telco Customer Churn](https://public.graphext.com/362cd8e3e96f70d4/index.html) | 7043 | 21 | Business | [Kaggle](https://www.kaggle.com/datasets/blastchar/telco-customer-churn) | | 22 | [Airbnbs Listings NY](https://public.graphext.com/77265ea3a63e650f/index.html) | 37012 | 33 | Travel and Locations | [Kaggle](https://www.kaggle.com/datasets/dgomonov/new-york-city-airbnb-open-data) | | 23 | [Climate in Madrid](https://public.graphext.com/83a75b4f1cea8df4/index.html?section=data) | 36858 | 26 | Travel and Locations | [AEMET](https://public.graphext.com/83a75b4f1cea8df4/index.html?section=data) | | 24 | [Salary Survey Spain 2018](https://public.graphext.com/24d1e717ba01aa3d/index.html) | 216726 | 29 | Business | [INE](ine.es) | | 25 | [Data Driven SEO ](https://public.graphext.com/4e5b1cac9ebdfa44/index.html) | 62 | 5 | Business | [Graphext](https://www.graphext.com/post/data-driven-seo-a-keyword-optimization-guide-using-web-scraping-co-occurrence-analysis-graphext-deepnote-adwords) | | 26 | [Predicting Wine Quality](https://public.graphext.com/de04acf5d18a9aea/index.html) | 1599 | 12 | Business | [Kaggle](https://www.kaggle.com/datasets/yasserh/wine-quality-dataset) | | 27 | [Supermarket Sales](https://public.graphext.com/9a6742da6a8d8f7f/index.html) | 1000 | 17 | Business | [Kaggle](https://www.kaggle.com/datasets/aungpyaeap/supermarket-sales) | | 28 | [Predict Diabetes](https://public.graphext.com/def4bada27af324c/index.html) | 768 | 9 | Health | [Kaggle](https://www.kaggle.com/datasets/iammustafatz/diabetes-prediction-dataset) | | 29 | [NYTimes World In 2021](https://public.graphext.com/af4c8eef1757973c/index.html?section=data) | 52588 | 5 | Travel and Locations | [New York Times](https://public.graphext.com/af4c8eef1757973c/index.html) | | 30 | [Professionals Kaggle Survey](https://public.graphext.com/3a2e87f90363a85d/index.html) | 19169 | 64 | Business | [Kaggle](https://www.kaggle.com/c/kaggle-survey-2021/data) | | 31 | [Trustpilot Reviews](https://public.graphext.com/367e29432331fbfd/index.html?section=data) | 8020 | 6 | Business | [TrustPilot](https://public.graphext.com/367e29432331fbfd/index.html?section=data) | | 32 | [Delicatessen Customers](https://public.graphext.com/a1687589fbde07bc/index.html) | 2240 | 29 | Business | [Kaggle](https://www.kaggle.com/datasets/rodsaldanha/arketing-campaign) | | 33 | [Employee Attrition](https://public.graphext.com/07a91a15ecf2b8f6/index.html) | 14999 | 11 | Business | [Kaggle(modified)](https://www.kaggle.com/datasets/pavan9065/predicting-employee-attrition) | | 34 | [World Happiness Report 2020](https://public.graphext.com/754c83ff0a7ba087/index.html) | 153 | 20 | Social Networks and Surveys | [World Happiness](https://worldhappiness.report/data/) | | 35 | [Billboard Lyrics](https://public.graphext.com/7e0b009e8d0af719/index.html) | 5100 | 6 | Sports and Entertainment | [Brown University](https://cs.brown.edu/courses/cs100/students/project11/) | | 36 | [US Migrations 2012-2016](https://public.graphext.com/dbdadf87a5c21695/index.html) | 288300 | 9 | Social Networks and Surveys | [US Census](https://www.census.gov/topics/population/migration/guidance/county-to-county-migration-flows.html) | | 37 | [Ted Talks](https://public.graphext.com/07e48466fb670904/index.html) | 4005 | 19 | Social Networks and Surveys | [Kaggle](https://www.kaggle.com/datasets/ashishjangra27/ted-talks) | | 38 | [Stroke Likelihood](https://public.graphext.com/20ccfee9e84948e3/index.html) | 5110 | 12 | Health | [Kaggle](https://www.kaggle.com/datasets/kamilpytlak/personal-key-indicators-of-heart-disease) | | 39 | [Happy Moments](https://public.graphext.com/9b86efff48989701/index.html) | 100535 | 11 | Social Networks and Surveys | [Kaggle](https://www.kaggle.com/datasets/ritresearch/happydb) | | 40 | [Speed Dating](https://public.graphext.com/f1912daad7870be0/index.html) | 8378 | 123 | Social Networks and Surveys | [Kaggle](https://www.kaggle.com/datasets/ulrikthygepedersen/speed-dating) | | 41 | [Airline Mentions X (former Twitter)](https://public.graphext.com/29cb7f73f6e17a38/index.html) | 14640 | 15 | Social Networks and Surveys | [X (former Twitter)](https://public.graphext.com/7e6999327d1f83fd/index.html) | | 42 | [Predict Student Performance](https://public.graphext.com/def4bada27af324c/index.html) | 395 | 33 | Business | [Kaggle](https://www.kaggle.com/datasets/impapan/student-performance-data-set) | | 43 | [Loan Defaults](https://public.graphext.com/0c7fb68ab8071a1f/index.html) | 83656 | 20 | Business | [SBA](https://www.kaggle.com/datasets/mirbektoktogaraev/should-this-loan-be-approved-or-denied) | | 44 | [IMDb Movies](https://public.graphext.com/e23e33774872c496/index.html) | 85855 | 22 | Sports and Entertainment | [Kaggle](https://www.kaggle.com/datasets/harshitshankhdhar/imdb-dataset-of-top-1000-movies-and-tv-shows) | | 45 | [Spotify Song Popularity](https://public.graphext.com/def4bada27af324c/index.html) | 21000 | 19 | Sports and Entertainment | [Spotify](https://www.kaggle.com/datasets/tomigelo/spotify-audio-features) | | 46 | [120 Years Olympics](https://public.graphext.com/e57d5e2f172c9a99/index.html) | 271116 | 15 | Sports and Entertainment | [Kaggle](https://www.kaggle.com/datasets/heesoo37/120-years-of-olympic-history-athletes-and-results) | | 47 | [Bank Customer Churn](https://public.graphext.com/e8f7aeacd209f74a/index.html) | 7088 | 15 | Business | [Kaggle](https://www.kaggle.com/datasets/mathchi/churn-for-bank-customers) | | 48 | [Data Science Salary Data](https://public.graphext.com/4e5b1cac9ebdfa44/index.html) | 742 | 28 | Business | [Kaggle](https://www.kaggle.com/datasets/ruchi798/data-science-job-salaries) | | 49 | [Boris Johnson UK PM Tweets](https://public.graphext.com/f6623a1ca0f41c8e/index.html) | 3220 | 34 | Social Networks and Surveys | [X (former Twitter)](https://public.graphext.com/f6623a1ca0f41c8e/index.html) | | 50 | [ING 2019 X Mentions](https://public.graphext.com/075030310aa702c6/index.html) | 7244 | 22 | Social Networks and Surveys | [X (former Twitter)](https://public.graphext.com/075030310aa702c6/index.html) | | 51 | [Pokemon Features](https://public.graphext.com/f30d4d863a2e6b01/index.html) | 1072 | 13 | Business | [Kaggle](https://www.kaggle.com/datasets/rounakbanik/pokemon) | | 52 | [Professional Map](https://public.graphext.com/70af2240cb751968/index.html) | 1227 | 12 | Business | [Kern et al, PNAS'20](https://github.com/behavioral-ds/VocationMap) | | 53 | [Google Patents](https://public.graphext.com/a262300e31874716/index.html) | 9999 | 20 | Business | [BigQuery](https://www.kaggle.com/datasets/bigquery/patents/data) | | 54 | [Joe Biden Tweets](https://public.graphext.com/33fa2efa41541ab1/index.html) | 491 | 34 | Social Networks and Surveys | [X (former Twitter)](https://public.graphext.com/339cee259f0a9b32/index.html?section=data) | 55 | [German Loans](https://public.graphext.com/d3f5e425e9d4b0a1/index.html) | 1000 | 18 | Business | [Kaggle](https://www.kaggle.com/datasets/uciml/german-credit/data) | | 56 | [Emoji Diet](https://public.graphext.com/e721cc7d790c06d4/index.html) | 58 | 35 | Health | [Kaggle](https://www.kaggle.com/datasets/ofrancisco/emoji-diet-nutritional-data-sr28) | | 57 | [Spain Survey 2015](https://public.graphext.com/90ca7539b160fdfa/index.html?section=data) | 20000 | 45 | Social Networks and Surveys | [CIS](https://public.graphext.com/90ca7539b160fdfa/index.html?section=data) | | 58 | [US Polls 2020](https://public.graphext.com/dbdadf87a5c21695/index.html) | 3523 | 52 | Social Networks and Surveys | [Brandwatch](https://www.brandwatch.com/p/us-election-raw-polling-data/) | | 59 | [Second Hand Cars](https://public.graphext.com/543d0c49d7120ca0/index.html) | 50000 | 21 | Business | [DataMarket](https://www.kaggle.com/datasets/datamarket/venta-de-coches) | | 60 | [Bakery Purchases](https://public.graphext.com/6f2102e80f47a192/index.html) | 20507 | 5 | Business | [Kaggle](https://www.kaggle.com/code/xvivancos/market-basket-analysis/report) | | 61 | [Disneyland Customer Reviews](https://public.graphext.com/b1037bb566b7b316/index.html) | 42656 | 6 | Travel and Locations | [Kaggle](https://www.kaggle.com/datasets/arushchillar/disneyland-reviews) | | 62 | [Trump Tweets](https://public.graphext.com/7aff94c3b7f159fc/index.html) | 15039 | 20 | Social Networks and Surveys | [X (former Twitter)](https://public.graphext.com/be903c098a90e46f/index.html?section=data) | | 63 | [Influencers](https://public.graphext.com/e097f1ea03d761a9/index.html) | 1039 | 14 | Social Networks and Surveys | [X (former Twitter)](https://public.graphext.com/e097f1ea03d761a9/index.html) | | 64 | [Clustering Zoo Animals](https://public.graphext.com/d1b66902e46a712a/index.html) | 101 | 18 | Health | [Kaggle](https://www.kaggle.com/datasets/jirkadaberger/zoo-animals) | | 65 | [RFM Analysis](https://public.graphext.com/4db2e54e29006a21/index.html) | 541909 | 8 | Business | [UCI ML](https://www.kaggle.com/datasets/carrie1/ecommerce-data) | ## 🏗️ Folder structure Each folder represents one dataset. You will find the following files within: * all.parquet: the processed data, with each column tagged with our typing system, in [parquet](https://arrow.apache.org/docs/python/parquet.html). * qa.csv: contains the human-made set of questions, tagged by type and columns used, for the dataset (sample_answer indicates the answers for DataBench lite) * sample.csv: sample containing 20 rows of the original dataset (DataBench lite) * info.yml: additional information about the dataset ## 🗂️ Column typing system In an effort to map the stage for later analysis, we have categorized the columns by type. This information allows us to segment different kinds of data so that we can subsequently analyze the model's behavior on each column type separately. All parquet files have been casted to their smallest viable data type using the open source [Lector](https://github.com/graphext/lector) reader. What this means is that in the data types we have more granular information that allows us to know if the column contains NaNs or not (following panda’s convention of Int vs int), as well as whether small numerical values contain negatives (Uint vs int) and their range. We also have dates with potential timezone information (although for now they’re all UTC), as well as information about categories’ cardinality coming from the arrow types. In the table below you can see all the data types assigned to each column, as well as the number of columns for each type. The most common data types are numbers and categories with 1336 columns of the total of 1615 included in DataBench. These are followed by some other more rare types as urls, booleans, dates or lists of elements. | Type | Columns | Example | | -------------- | ------- | ----------------------- | | number | 788 | 55 | | category | 548 | apple | | date | 50 | 1970-01-01 | | text | 46 | A red fox ran... | | url | 31 | google.com | | boolean | 18 | True | | list[number] | 14 | [1,2,3] | | list[category] | 112 | [apple, orange, banana] | | list[url] | 8 | [google.com, apple.com] | ## 🔗 Reference You can download the paper [here](https://huggingface.co/datasets/cardiffnlp/databench/resolve/main/Databench-LREC-Coling-2024.pdf). If you use this resource, please use the following reference: ``` @inproceedings{oses-etal-2024-databench, title = "Question Answering over Tabular Data with DataBench: A Large-Scale Empirical Evaluation of LLMs", author = "Jorge Jorge Osés Grijalba and Luis Alfonso Ureña-López and Eugenio Martínez Cámara and Jose Camacho-Collados", booktitle = "Proceedings of LREC-COLING 2024", year = "2024", address = "Turin, Italy" } ```