Summary
A leading Canadian bank possessed a large model data set from disparate sources. Evalueserve cleaned, standardized, and transformed the data set, improving the bank’s data architecture and ensuring full regulatory compliance. Evalueserve’s efforts resulted in a shorter model development process and a data repository that was applicable to the broad risk landscape.
The Problem
The Canadian bank had a massive set of model data and a large suite of complex products. Their data came from several sources and lacked standardization. Before their data could be used to build regulatory models, it needed to be cleaned, analyzed, and transformed for model development.
The bank hired Evalueserve to aggregate disparate data sources to create a single source of truth for large data pipelines for model development, monitoring, and validation. They also wanted to improve data quality and consistency across current and future data sources.
Our Solution
Evalueserve improved the client’s data architecture by building a data repository.
First, Evalueserve’s experts aggregated and consolidated data from disparate sources. Next, they automated the data feed to clean, ingest, and enhance that data and transform independent variables into the right variables, so the entire analysis process was faster. The automated data feed and analysis can integrate with business intelligence (BI) tools.
Then, the team created documentation around mapping data from the source to the target database. Lastly, we put consolidated data rules in place and determined Critical Data Elements (CDEs) that are monitored every month to ensure data consistency and quality.
Business Impact
The client was provided with a single source of truth thanks to Evalueserve’s cleaning, analysis, and transformation efforts. As a result, the Canadian bank now has a data repository with high-quality data that applies to the broad risk landscape, from credit and market risk to counterparty and operational/enterprise risk.
Usually, around 40% of the entire model development process is spent on the data cleaning, preparation, and transformation needed before data can be used for model fitting. Evalueserve’s experts significantly reduced the time required to clean and prepare data.
The documentation of the data process increased visibility and transparency across the model lifecycle, which saved the model validation team significant time.
Learn more about Evalueserve’s financial risk solutions.