From its incorporation, OSIS has been very active in providing data management solutions to financial institutions. Structured data is an important cornerstone for all analysis and statistical modelling for any lender or investor in loans.
OSIS offers data management solutions to help banks, Fintech companies, asset managers, pension funds, and insurance companies effectively manage their credit risk. Our solutions include powerful tools for data collection, storage, and analysis, as well as expert advice and training to help you develop the expertise needed to manage your data effectively.
Structured and unstructured data
Structured data are grouped in fixed fields following a predefined data model. Examples for structured data are birth dates, addresses, account numbers, industry sector, legal form, and default dates. Unstructured data don’t follow a data model and are usually stored in their original format. Typical examples of unstructured data are meeting minutes, credit reports, photographs, and social media activity. Unstructured data can’t be processed with conventional tools and methods. New techniques to mine and arrange unstructured data are still in the development phase and are therefore a challenge. About 80% of all the data in companies is unstructured.
Advantages of structured data
Structured data have many advantages in comparison to unstructured data. They can be (1) processed automatically in huge quantities with (2) matured techniques. Further, if the data follow a well-defined data model (3) the quality can be automatically assessed, and (4) erroneous fields can be identified. Further, (5) statistical models can be calibrated in the simplest and most intuitive way. If the same data model is used across different companies their (6) data can be more easily analysed and compared by their respective stakeholders and be pooled with the data from other lenders forming a (7) larger statistical significance leading to better models.
Challenges with structured data
Structured data are grouped in fixed fields following a predefined data model. Examples for structured data are birth dates, addresses, account numbers, industry sector, legal form, and default dates. Unstructured data don’t follow a data model and are usually stored in their original format. Typical examples of unstructured data are meeting minutes, credit reports, photographs, and social media activity. Unstructured data can’t be processed with conventional tools and methods. New techniques to mine and arrange unstructured data are still in the development phase and are therefore a challenge. About 80% of all the data in companies is unstructured.
Data curation is what we offer
Data curation is the (1) organisation and (2) integration of data collected from various sources. It involves annotation, publication, and presentation of the data such that the value of the data is (3) maintained over time, and the data remains (4) available for reuse and preservation. Data curation includes “all the processes needed for principled and controlled data creation, maintenance, and management, together with the capacity to add value to data”.