Case Study

Human-Augmented Data Curation & Validation


Our client’s unstructured financial transaction data was programmatically categorized at a high level but was not as granular as the client required. Human review was needed to provide additional context.


Best Practices

Data Type

Semi-Structured Text (.CSV, .PDF, .HTML, .XML)

Project Duration

2 Months




Associates cleaned and mapped strings of data related to transactions. Many transactions listed two or more businesses; without a human in the loop, it was difficult to identify to which entity the money from the transaction was going. Associates used context and independent research to identify how to correctly map the transaction to the associated company.


DataInFormation successfully maintains workflows to categorize and clean the provided data in a more nuanced way than possible with AI alone. This work supports the client’s tracking and categorization of metrics such as consumer trends, third-party sales, subsidiary impact, and more.

Download Case Study