Tellimer Group (Tellimer), the London-headquartered financial technology, data and information provider, has launched its proprietary data extraction and structuring tool, Parsel.
Read More: Gemini selects ClearBank as UK banking services provider
“Parsel is our first truly dedicated technology product and we believe we’ve built something ground-breaking. Combining technologies at the vanguard of machine-learning and optical character recognition, we expect Parsel to become an essential workflow tool for data professionals everywhere.”
As big data collection and management continues to inundate and challenge businesses around the world, the need to efficiently recover information from large, unstructured datasets is paramount. Using advanced AI algorithms and the latest optical character recognition technology, Parsel analyses tabular data trapped in PDFs and other image files and produces easy-to-read, editable tables of structured information in minutes, in a range of file formats including Excel (XLSX), Comma-Separated Values (CSV), and JavaScript Object Notation (JSON).
Uniquely, Parsel’s algorithms require no prior training on datasets and are benchmarked at an industry-leading 96.6% accuracy on financial documents — with an average of 91.6% across all tested documents — saving users a great deal of time in spot checks and corrections. Parsel also utilises the latest data security practices as standard, encrypting data both in-transit and at-rest, storing all information in a fully secure cloud environment.
Duncan Wales, CEO of Tellimer Group, said: “Parsel equips users with the ability to realise the hidden value in their unstructured data, and was created with analysts and finance professionals in mind. Whether it is company financial reports, financial and legal documentation or colossal macroeconomic datasets trapped in PDF documents, Parsel will improve business efficiency and save users time, effort and money when working with big data.”
Read More: SAP Extends Its Lead in the Dow Jones Sustainability Indices
