NIBC Case: Performing Analyses Faster through Machine Learning

NIBC Bank and Young Mavericks started a text data mining project at the Agency Management and Operations departments. Text mining is a form of analysis in which relevant information is extracted from large amounts of text material. The solution supports the banks’ employees to provide clients with accurate and relevant information, almost in real-time. By automating document review processes, employees can focus on offering personalized customer services to their clients and leave the routine analyses to machines. The application is ready for implementation.

Personalized Information for Each Department

The Agency Management department is responsible for the communication between the borrower (loan recipient) and the lenders (the syndicate of banks that provide the loan). The Operations department is responsible for administrative tasks and operational loan processing. Both departments work on documents that can span 700+ pages and document reviews can therefore take hours or sometimes even days to complete per department. 

Huge efficiency gains can be achieved by automatizing part of the work. After a month, there was plenty of prove that the technology could help employees do their work accurate and much faster.

Get Your Information Faster with a Text Mining Application

Duncan de Vries, Innovation Manager at NIBC, and Wim van Velzen, Head of Agency Management, partnered with Young Mavericks to speed up the cumbersome text data reviews. Young Mavericks’ work is focused on future-proofing the ties between organizations and talent. We accomplish this goal by deploying the most important and recent technological developments in Data Engineering, Data Science and UX Design to both business and non-profit programs and projects. NIBC and Young Mavericks worked together to solve the departments’ laborious review time by passing the documents through a customized text mining application. The code for the text mining application has been written to scan, identify and filter the relevant information automatically from the voluminous documents.

The Results

The initial KPI (Key Performance Indicator) was to determine the possibility of extracting specific events from contracts resulting in an accuracy rate well above 80%. Two different types of events and more than twenty administration details were automatically extracted from the contracts. This KPI was met within a couple of weeks.

Two new KPI’s were formulated after the initial KPI was achieved; the accuracy and efficiency of the model. An important aspect to take into account with the model’s accuracy is that, just like an employee’s work performance, a machine learning model will unlikely achieve a 100% score. A successful application does not have to be perfect, it has to be better and faster than a human being. Additionally, the combination of computer and employee will result in a better performance. 

The model now correctly identifies 89-95% of the events and >98% of all administration details. The efficiency KPI was easily met, as work that normally takes several hours can now take place in a time span of a few seconds by combining text mining and machine learning technologies. A huge gain of time!

Duncan de Vries describes the effort:

“This machine learning proof of concept showed its usefulness within a month. We can perform our analyses much faster for these specific use cases. The proof of concept also inspired other employees, which led to several other text mining use cases. We are looking forward to expanding this case and to deploy the technology more widely in the organization.”

How Docker and Kubernetes’ Open Source Technology is Winning Over Businesses

By Young Mavericks’ Data Engineer Don de Lange

Data Engineers must ensure that technological solutions for companies can actually be implemented. In order to fulfil these technical promises in an increasingly complex data-driven world, like many other Data Engineers, I use various software and data tools. Docker and Kubernetes, two leaders in the field of open source technologies, help to build, manage and scale apps in containers. In this blog I explain what Docker and Kubernetes are, why more and more companies are taking advantage of their expertise and how you yourself can take advantage of these platforms.

Read more

Interview with Data Engineer Joeri: About solving puzzles, tadpoles and the language of programming

Young Mavericks’ Data Engineer Joeri (26) helps professionalize the Waterschap (Water Board) Aa en Maas. He is working on ‘Python Refactoring’ in their Datalab, a process aimed at optimizing existing codes, which makes them more accessible, user-friendly, easier to manage and more error-proof. Joeri: “Colleagues at Aa and Maas did not immediately see the added value of refactoring. By showing the time and money it could save them, I was able to win them over. Now, everybody is excited.”

Read more