Mathnasium Solves its Data Analysis Equation with Adastra and AWS

Mathnasium, a franchise-based, math-tutoring organization with over 1,100 locations around the world, was looking to better understand its data and simplify analytics. Struggling with a legacy database system, it turned to AWS Advanced Partner Adastra, to optimize its data analytics. Using several AWS services, Adastra migrated data from SQL Server on Amazon RDS to Amazon S3 and enabled Amazon QuickSight for business intelligence and the ability to create dashboards. As a result, Mathnasium gained critical insights into the performance of its franchise locations that now inform its overall business strategy.

Given that Mathnasium is literally built on numbers, the leadership team was frustrated they couldn’t obtain their performance indicators (KPIs) faster. Mathnasium’s legacy system included a single database, and this posed two key challenges—slow and inefficient report generation and difficulty in updating the system. The inefficiency was because the same database was used for both report generation and writing or updating transactions. Any updates or changes made to the system impacted day-to-day business. Additionally, data was not aggregated optimally. These two factors slowed the metric reporting process considerably for development teams.

From a business perspective, the system was unable to effectively generate key metrics. Jason Condello, chief technology officer at Mathnasium, explained, “We wanted to understand the business potential of each student, what their tenure was, their progress between their pre- and post-program assessments, and the student-instructor ratio at our tutoring centers.” Without these KPIs, Mathnasium headquarters was unable to see which of its franchises was under-performing or in need of support.

Already running on Amazon Web Services (AWS), Mathnasium chose to team up with AWS Partner Adastra, who specializes in delivering fast analytics on the cloud and providing a suite of AWS services around data estate modernization, analytics, artificial intelligence, governance, and managed services. Working together, Adastra developed a new data and analytics architecture for Mathnasium for using a data lake and data warehouse.

This entailed migrating data from Mathnasium’s legacy system stored on Amazon Relational Database Service (Amazon RDS) to Amazon Simple Storage Service (Amazon S3) to allow better data access and query response. Data pipelines built using AWS Glue, AWS Step Functions, Amazon Dynamo DB, and AWS Lambda enabled the move. In addition, AWS Database Migration Service (AWS DMS) provided change data capture (CDC) for data from SQL Server, ensuring Mathnasium brought over the latest data updates. A data warehouse built using Amazon Redshift and a data lake built on Amazon S3 completed the solution architecture. Finally, Amazon Redshift datasets were created to be used by Amazon QuickSight, offering access to powerful data analysis capabilities. The process of solution development and deployment was easy thanks to the native integration of AWS services. This enabled the democratization of data analysis access across Mathnasium, including at franchise locations, and changed the way the organization informs its business strategy.

The Mathnasium team is delighted with the improved documentation and reporting. As Condello put it, “We now have the ability to measure things we could not even see before.” This level of accessibility and transparency in analytics, complemented by the ease of deployment and flexibility to scale capacity, has proved a winner for Mathnasium. The true benefits have been the detailed metrics that Mathnasium and its franchisees can now access, a smooth deployment process thanks to robust continuous integration, continuous delivery (CI/CD) pipelines, and rapid data processing.

As Mathnasium migrated its historical data over to Amazon RDS, the team took advantage of on-demand compute on AWS and pay-as-you-go pricing to accelerate the process. The team first loaded data into the development environment, then again into the production environment. It was a massive lift that involved moving data that went back to the foundation of Mathnasium. “Being on the cloud let us crank up the server to 20 times its regular speed for a short period of time,” Condello said. “Once we had processed all the data, we simply turned it back down. And it cost us way less to do that than conventional methods.”

With a better handle on key metrics, Condello envisages building franchisee dashboards for each of its locations. The foundational architecture that Adastra put in place ensures that Mathnasium is future ready—with pipelines for machine learning models that make the analytics process more efficient. This will make it easier to innovate moving forward and integrate with additional AWS services. “Each AWS service acts like a Lego block,” said Elkin Arboleda, AWS practice lead at Adastra. “You need very little coding to use them, and they play nicely with each other.” Clearly, the future is adding up in neat stacks for Mathnasium.