Headquartered in Pontiac, Michigan, United Wholesale Mortgage provides mortgage products and services to mortgage brokers all over the country and is currently the top wholesale and mortgage lender in the United States.

In a large IT organization such as UWM, there are thousands of software changes and custom software solutions built every year. These changes produce valuable data about how certain software changes can affect the overall development cycle and other production risks. However, the data associated with these changes is spread across multiple different systems and is ineffective in this state.

Our Change Insights Datamart and Risk Assessment tool assists team leaders at UWM to proactively mitigate any potential production risks throughout the development cycle and monitor their team’s performance.

In our tool, data related to software changes is aggregated from various sources into one cohesive IT Datamart. This creates a single view of all IT operations within UWM, making it easier than ever to analyze data at a glance.

Our predictive model leverages data from the IT Datamart to collect crucial insights which may correlate to deployment risks such as net changes to a file, associated incident reports, and which team is contributing the corresponding changes. Following collection and analyzation, the model determines the level of risk associated with each software change.

Our IT Datamart includes data from Bitbucket, Jira, Harness, Octopus, and ServiceNow. Python scripts are utilized for cleaning the CSV files. The model is implemented in Azure Machine Learning Studio. Azure Blob Storage is used to import data into Power BI from the model for front-end representation.