Cognaptus DataHub Resource Index
Curated datasets and publicly listed model resources for business process automation discovery.
BPA Datasets
Curated data sources for document AI, workflow automation, and enterprise chatbot development.
Receipt Dataset for Document Forgery Detection
Relevant because it matches: receipt, document, business.
Walmart Sales Forecasting A CRISP-DM Model
Relevant because it matches: business, sales, retail, forecast.
A document annotation dataset to perform NER on resumes.
Relevant because it matches: document, ner, business.
Five years of daily retail sales (2019-2023)
Relevant because it matches: business, sales, retail, forecasting.
Curated Dataset of Invoice Images for Training and Evaluation
Relevant because it matches: invoice, ocr, business.
Extracting text from a transaction receipt using OpenCv
Relevant because it matches: receipt, business, transaction.
Explore Open Data from World Bank Financial
Relevant because it matches: contract, finance, procurement, business.
Valan Procurement Intelligence - Public Sample A free 1,000-row sample of the Valan Technologies global procurement intelligence feed: point-in-time, entity-resolved government contract awards linked to the awarded company's tradable security (ticker / LEI / ISIN). Valan Technologies maintains one of the most comprehensive privately-held public procurement datasets in existence: 100M+ contract awards and tenders from 150+ government sources across 222 countries and territories... See the full description on the dataset page: https://huggingface.co/datasets/ValanTec/procurement-sample.
Relevant because it matches: contract, finance, procurement.
RetailSalesAnalyzer tags: consumer behavior, sales trend, segmentation Note: This is an AI-generated dataset so its content may be inaccurate or false Dataset Description: RetailSalesAnalyzer is a comprehensive dataset capturing consumer purchasing patterns and retail performance metrics. The dataset includes detailed transaction records with customer demographics, product categories, seasonal trends, and behavioral segmentation data. It enables analysis of buying habits, market... See the full description on the dataset page: https://huggingface.co/datasets/infinite-dataset-hub/RetailSalesAnalyzer.
Relevant because it matches: records, sales, retail, customer. Watch-outs: synthetic.
Multimodal, rule-explained synthetic/semi-synthetic procurement invoice fraud DS
Relevant because it matches: invoice, finance, procurement, fraud. Watch-outs: synthetic.
Photos of the receipts and text detection - ocr dataset
Relevant because it matches: ocr, business, retail.
This only a subpart of the original dataset, containing only invoice. The RVL-CDIP (Ryerson Vision Lab Complex Document Information Processing) dataset consists of 400,000 grayscale images in 16 classes, with 25,000 images per class. There are 320,000 training images, 40,000 validation images, and 40,000 test images. The images are sized so their largest dimension does not exceed 1000 pixels. For questions and comments please contact Adam Harley ([email protected]). The full dataset... See the full description on the dataset page: https://huggingface.co/datasets/chainyo/rvl-cdip-invoice.
Relevant because it matches: invoice, document.
A Detailed Analysis of Global Car Sales Trends
Relevant because it matches: report, business, sales.
EDA and Analysis Ideas for restaurant business
Relevant because it matches: report, business, sales.
Detailed records of customer support tickets
Relevant because it matches: records, retail, customer, support.
Uncover Patterns and Trends for Enhanced Sales Strategies
Relevant because it matches: crm, business, sales.
Dataset for training Triage Bots and Sentiment Analyzers
Relevant because it matches: crm, customer, support.
Learn time series forecasting with yahoo finance stock price
Relevant because it matches: finance, business, forecasting.
Focused customer retention programs
Relevant because it matches: business, customer, churn.
A suitable data for CRM & Sales Opportunities & Marketing analytics practice.
Relevant because it matches: crm, sales, retail.
A suitable data for CRM & Sales Opportunities & Marketing analytics practice.
Relevant because it matches: crm, business, sales.
Retail Trends: A Deep Dive into Transactional Data
Relevant because it matches: finance, business, retail, transaction.
Sales transaction of a UK-based e-commerce (online retail) for one year
Relevant because it matches: business, sales, retail, transaction.
The place for Gold Future prices (GC00: Gold Continuous Contract Futures, USD)
Relevant because it matches: contract, finance.
A collection of ~1000 newsgroup documents from 10 different newsgroups
Relevant because it matches: document, business.
Hand labelled corpus of legal contracts, courtesy of atticusprojectai.org
Relevant because it matches: contract, business.
A real online retail transaction data set of two years
Relevant because it matches: business, retail, transaction.
Identifying and Preventing Fraudulent Activities in Financial Transactions
Relevant because it matches: finance, fraud, transaction.
A real online retail transaction data set of two years.
Relevant because it matches: business, retail, transaction.
Customer Behavior and Engagement Dataset for Churn Modeling
Relevant because it matches: business, retail, customer, churn.
1M transactions . 7 fraud types . fraud rings . behavioural features . class imb
Relevant because it matches: finance, insurance, fraud.
"The moment you make a mistake in pricing, you're eating into your reputation"
Relevant because it matches: business, retail, pricing.
Daily COMEX Gold Futures (GC) data curated for quantitative research, time serie
Relevant because it matches: contract, finance.
Dataset Card for Financial Fraud Labeled Dataset Dataset Details This dataset collects financial filings from various companies submitted to the U.S. Securities and Exchange Commission (SEC). The dataset consists of 85 companies involved in fraudulent cases and an equal number of companies not involved in fraudulent activities. The Fillings column includes information such as the company's MD&A, and financial statement over the years the company stated on the SEC... See the full description on the dataset page: https://huggingface.co/datasets/amitkedia/Financial-Fraud-Dataset.
Relevant because it matches: finance, fraud.
Dataset Card for Thai Government Procurement Dataset i This dataset is optimized for procurement-related NLP tasks in Thai. This dataset contains a collection of procurement regulations, instructions, and responses focused on public sector purchasing, contract management, and compliance with Thai government standards. It aims to support natural language processing tasks involving procurement assistance, such as chatbot development, procurement dialogue generation... See the full description on the dataset page: https://huggingface.co/datasets/amornpan/thai-gov-procurement_regulation-17-amend-21.
Relevant because it matches: contract, procurement, support. Watch-outs: sft.
Purchase History Records | Africa (Electric Sheep Africa metadata inventory) Size category: 1M<n<10M - Formats: parquet - Sector: culture_language - Engineered by Electric Sheep Africa TL;DR This dataset is part of the Electric Sheep Africa catalog on Hugging Face. It is indexed for African data discovery with standardized metadata, loading guidance, provenance notes, and analyst-oriented context. What This Dataset Covers Public datasets help... See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-synth-retail-and-ecommerce-purchase-history-records-nigeria.
Relevant because it matches: records, sales, retail, ecommerce. Watch-outs: synthetic.
customer-support-th-26.9k Thai customer-support instruction dataset (~26.9k examples). Thai-localized version of the Bitext customer-support dataset - instruction templates, intent/category labels, and response templates in Thai. Format Field Description instruction Customer question template in Thai (may contain {{placeholders}}) response Support response template in Thai category Coarse category (e.g. ORDER) intent Fine-grained intent (e.g.... See the full description on the dataset page: https://huggingface.co/datasets/Porameht/customer-support-th-26.9k.
Relevant because it matches: customer, support.
Dataset Card for Dataset Name This dataset card aims to be a base template for new datasets. It has been generated using this raw template. Dataset Details Dataset Description Curated by: [More Information Needed] Funded by [optional]: [More Information Needed] Shared by [optional]: [More Information Needed] Language(s) (NLP): [More Information Needed] License: [More Information Needed] Dataset Sources [optional] Repository: [More... See the full description on the dataset page: https://huggingface.co/datasets/RachidAb02/Finance-Accounting.
Relevant because it matches: finance, accounting.
LexMetrix AltData: Dominican Procurement Risk Intelligence (2021-2026) Public procurement dataset + risk scoring model for monopoly detection, compliance auditing, and open-data research across Latin America. Dataset Overview Metric Value Records 305,094 government contracts Period January 2021 - June 2026 Government spending analyzed USD 14.6 billion CRITICAL risk flags 2,632 contracts (0.86%) 5+ year monopoly patterns 193 contracts... See the full description on the dataset page: https://huggingface.co/datasets/EmmanuelJDP/dgcp-procurement-risk-do.
Relevant because it matches: records, procurement.
Ecommerce FAQ Chatbot Dataset Overview The Ecommerce FAQ Chatbot Dataset is a valuable collection of questions and corresponding answers, meticulously curated for training and evaluating chatbot models in the context of an Ecommerce environment. This dataset is designed to assist developers, researchers, and data scientists in building effective chatbots that can handle customer inquiries related to an Ecommerce platform. Contents The dataset comprises a total of 79 question-answer pairs... See the full description on the dataset page: https://huggingface.co/datasets/Andyrasika/Ecommerce_FAQ.
Relevant because it matches: ecommerce, customer.
cik form accessionNumber fileNumber filmNumber reportDate url 1463814 ABS-EE 0000950131-18-001228 333-212311 18946066 2018-06-30 https://sec.gov/Archives/edgar/data/1463814/000095013118001228 1463814 ABS-EE 0000950131-18-001229 333-212311 18946077 2018-06-30 https://sec.gov/Archives/edgar/data/1463814/000095013118001229 1463814 ABS-EE 0000950131-18-001670 333-212311 181029607 2018-07-31 https://sec.gov/Archives/edgar/data/1463814/000095013118001670 1463814 ABS-EE... See the full description on the dataset page: https://huggingface.co/datasets/DenyTranDFW/Mercedes_Benz_Retail_Receivables_LLC_1463814.
Relevant because it matches: form.
Medium High Level Implementation Of Sustainable Public Procurement | Europe (Our World in Data) 42 observations . 26 Europe countries . 2020-2022 . Repackaged by Electric Sheep Europe TL;DR This dataset contains 42 observations of Medium High Level Implementation Of Sustainable Public Procurement data across 26 Europe countries, spanning 2020-2022. About the source Source: Our World in Data Publisher: Our World in Data License: cc-by-4.0... See the full description on the dataset page: https://huggingface.co/datasets/electricsheepeurope/europe-owid-medium-high-level-implementation-of-sustainable-public-procurement.
Relevant because it matches: procurement, forecasting.
FinanceQA is a comprehensive testing suite designed to evaluate LLMs' performance on complex financial analysis tasks that mirror real-world investment work. The dataset aims to be substantially more challenging and practical than existing financial benchmarks, focusing on tasks that require precise calculations and professional judgment. Paper: https://arxiv.org/abs/2501.18062 Description The dataset contains two main categories of questions: Tactical Questions: Questions based on... See the full description on the dataset page: https://huggingface.co/datasets/AfterQuery/FinanceQA.
Relevant because it matches: finance.
Embeddings derived from business descriptions of S&P500 companies using sentence-BERT, SentenceTransformer('all-MiniLM-L6-v2') to be exact. For more info on evaluation of sentence transformers (specifcailly the huge GPT-3 versus smaller models see: https://twitter.com/Nils_Reimers/status/1487014195568775173)
Relevant because it matches: business.
Dataset Card for "ads-retail" More Information needed
Relevant because it matches: retail.
This dataset is a combination of Stanford's Alpaca (https://github.com/tatsu-lab/stanford_alpaca) and FiQA (https://sites.google.com/view/fiqa/) with another 1.3k pairs custom generated using GPT3.5 Script for tuning through Kaggle's (https://www.kaggle.com) free resources using PEFT/LoRa: https://www.kaggle.com/code/gbhacker23/wealth-alpaca-lora GitHub repo with performance analyses, training and data generation scripts, and inference notebooks: https://github.com/gaurangbharti1/wealth-alpaca... See the full description on the dataset page: https://huggingface.co/datasets/gbharti/finance-alpaca.
Relevant because it matches: finance.
Dataset Card for "contract" More Information needed
Relevant because it matches: contract.
Supply Chain Logistics Data | Africa (Electric Sheep Africa metadata inventory) Size category: 100K<n<1M - Formats: parquet - Sector: culture_language - Engineered by Electric Sheep Africa TL;DR This dataset is part of the Electric Sheep Africa catalog on Hugging Face. It is indexed for African data discovery with standardized metadata, loading guidance, provenance notes, and analyst-oriented context. What This Dataset Covers Public datasets... See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-synth-retail-and-ecommerce-supply-chain-logistics-data-nigeria.
Relevant because it matches: logistics, retail, ecommerce. Watch-outs: synthetic.
Dataset Card for "ecommerce-faq-chatbot-dataset" More Information needed
Relevant because it matches: ecommerce.
BPA Model Resources
Public listings are descriptive discovery artifacts, not deployment, security, commercial-use, or license-compatibility conclusions.
Public Kaggle and Hugging Face model resources ranked for BPA relevance. A listed license is disclosure, not a Cognaptus legal compatibility determination.
impira/layoutlm-document-qa
Positive: document; pdf; pipeline:document-question-answering
qwen-lm/qwen-3-vl
Qwen3-VL is the most powerful vision-language model in its series, offering significant upgrades in visual perception, reasoning, and context understanding
Positive: ocr; document; demand
impira/layoutlm-invoices
Positive: document; pdf; pipeline:document-question-answering
jinhybr/OCR-DocVQA-Donut
Positive: ocr; document; pipeline:document-question-answering
cloudqi/CQI_Visual_Question_Awnser_PT_v0
Positive: document; pdf; pipeline:document-question-answering
faisalraza/layoutlm-invoices
Positive: document; pdf; pipeline:document-question-answering
pranithkantamneni/frauddetection
Positive: form; fraud; transaction
google/embeddinggemma
EmbeddingGemma is a 300m parameter, state-of-the-art for its size, open embedding model from Google, built from Gemma 3 (with T5Gemma initialization) and the same research and technology used to create Gemini models.
Positive: document; form; pdf; anomaly
Negative: benchmark
serhiikharchuk/llama-3.1-minitron-4b-width-base
Llama 3.1 Minitron 4B Width Base
Positive: document; pdf
ahmedsayed0007/ask-your-pdf-rag-system
Positive: document; pdf
ashrithraojupelly/sentiment-analysis-and-clusters
Positive: document; form
golammostofas/astra
ASTRA LLM: Evokes vast knowledge, like the stars or the universe.
Positive: document; form
eachadea/ggml-vicuna-13b-1.1
Positive: document; pipeline:document-question-answering
naver-clova-ix/donut-base-finetuned-docvqa
Positive: document; pipeline:document-question-answering
fxmarty/tiny-doc-qa-vision-encoder-decoder
Positive: document; pipeline:document-question-answering
tiennvcs/layoutlmv2-base-uncased-finetuned-docvqa
Positive: document; pipeline:document-question-answering
rubentito/layoutlmv3-base-mpdocvqa
Positive: document; pipeline:document-question-answering
lakshminarasimhau/land-price-prediction-data-statistical-methods
Positive: document; forecasting
MariaK/layoutlmv2-base-uncased_finetuned_docvqa_v2
Positive: document; pipeline:document-question-answering
alimohamed01/named-entity-recognition-ner-v2
Positive: form; ner
tiennvcs/layoutlmv2-large-uncased-finetuned-infovqa
Positive: document; pipeline:document-question-answering
KyleDeanML/MNIME-Core-1.5B-Q4_K_M
Positive: document; pipeline:document-question-answering
KyleDeanML/MNIME-Core-V5-Q4_K_M
Positive: document; pipeline:document-question-answering
tiennvcs/layoutlmv2-base-uncased-finetuned-infovqa
Positive: document; pipeline:document-question-answering
angelchaudhary/credit-card-fraud-detection-model
Exploring machine learning techniques for accurate fraud detection using a large dataset and robust evaluation metrics.
Positive: fraud; transaction
andrometocs/andro_lstm_rnn_deeplearning_model_v1
Positive: forecast; forecasting; anomaly; demand
googleai/timesfm
The pretrained time-series foundation model developed by Google Research for time-series forecasting.
Positive: forecast; forecasting
payaldhokane/fraud-sense
Advancing Fraud Detection through Tabular GAN Based Synthetic Transactions
Positive: fraud; transaction
hassanjameelahmed/global-corn-supply-storytelling
Positive: forecasting; demand
csmohamedayman/ml-reg-tsf-store-item-demand-forecasting
Positive: forecasting; demand
shriyasingh900/walmart-project
Sales from Walmart, used for analyzing trends, forecasting demand, and optimizing retail operations.
Positive: forecast; forecasting; demand; transaction
krishnagarigipati/credit-card-fraud-detector-xgboost
A high-accuracy model for detecting fraudulent transactions with optimized performance.
Positive: fraud; transaction
mirajuddin1357/credit-card-fraud-detection-classification
Positive: fraud; transaction
google/gemma-3
Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models.
Positive: document; pdf
Negative: benchmark
google/t5gemma
T5Gemma is a family of encoder-decoder large language models, developed by adapting pretrained decoder-only Gemma into encoder-decoder.
Positive: document; pdf
Negative: benchmark
google/vaultgemma
VaultGemma is a variant of the Gemma family of lightweight, state-of-the-art open models from Google
Positive: document; pdf
Negative: benchmark
metaresearch/llama-3.2-vision
The Llama 3.2-Vision collection of multimodal large language models (LLMs) is a collection of pretrained and instruction-tuned image reasoning generative models in 11B and 90B sizes (text + images in / text out).
Positive: document; contract; pdf
Negative: benchmark
google/t5gemma-2
T5Gemma 2 is the next evolution of T5Gemma, adapted based on Gemma 3, featuring the first multi-modal and long-context encoder-decoder models.
Positive: document; pdf
Negative: benchmark
qwen-lm/qwen2.5-vl
Qwen2.5-VL is the new flagship vision-language model of Qwen
Positive: ocr; document
Negative: benchmark
domdejonge/essay-detection
a set of models used to detect if an essay was written by humans or Large Language models
Positive: document; form
Negative: medical