Sign in to view source links and access this dataset
Description
VynFi provides 2,106,112 journal entry line items from the manufacturing sector. The data originates from 10 companies over 12 monthly periods and includes labels for revenue fraud, vendor kickback, payroll ghost, and management override. It was last updated on the Hugging Face platform in April 2026.
Use Cases
Train fraud classification models based on the 6.92% fraud rate and specific fraud type labels.
Analyze patterns in manual journal entries based on the 21% manual entry rate relevant to ISA 240 auditing standards.
Study double-entry accounting mechanics based on the description of debit and credit postings.
Conduct sector-specific financial analysis based on data from the manufacturing industry.
Strengths
Large scale with 2,106,112 line items from 200,384 documents.
Includes explicit fraud labels covering four specific types at a 6.92% rate.
Sector-specific data from 10 manufacturing companies over 12 months.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
VynFi, via Hugging Face.
Collection Method
Journal entry line items from 10 manufacturing companies, with parent document header fields merged in.
Time Range
Covers 12 monthly periods (specific years not stated).
Freshness
Last updated 2026-04-14 17:36:38; freshness should be verified.
License information is unknown and should be verified before use.