SBIA and Studio 3T
That's when we came across Studio 3T and we found it very helpful in terms of how to actually learn mongoshell, process it and view the data.
Michael Bryce Herrington, Research/COO, SBIA
SBIA – Sports Betting Innovative Analytics – is an analytics business delivering insights into sports performance.
Created in the wake of the overturning of PASPA in 2018, SBIA has sought to apply modern analytics and machine learning to sports data.
The process involves ingesting large amounts of sports data and processing it through applications and models crafted by the company’s team of data scientists, database managers, and traditional sports handicappers.
This in turn involves MongoDB Atlas databases and AWS applications run in an EC2 cluster, orchestrated by Apache Airflow. Workflows are defined in Python, which is also the core language used by the SBIA team.
The team has to be able to view, query, and create aggregations on the MongoDB data which can be quickly converted to Python code for use in their workflows.
Additionally, new joiners to the company traditionally have SQL rather than MongoDB skills and the company needs effective tools to reskill them and bring them quickly up to a productive level in this new environment.
Studio 3T’s GUI and IDE provides powerful tools for aggregation development, reskilling and onboarding new colleagues.
1. Aggregations can be automatically converted into Python code for use in the process workflow by Studio 3T and used in live updates.
2. The Aggregation Editor’s split view of input and output data for each individual stage is a really powerful way to improve the team’s productivity by enabling rapid iteration on query performance-tuning cycles.
3. Quickly accessible views of data can be created and managed in the 3T GUI-IDE, allowing everyone on the team to monitor and track performance of their multiple workflows and models.
4. Users with SQL skills can straightway use the SQL query tool to learn how SQL translates into MongoDB’s aggregation code, speeding up their transition to writing MongoDB aggregation pipelines natively. Being able to onboard new team members by rapidly translating their existing skills into MongoDB specific know-how is highly valued across the business.
The Benefits of Studio 3T
On moving to Studio 3T
“We built a team of different data scientists, database managers, and traditional handicappers.”, explains Michael Bryce Herrington – Research/COO.
”And back then we made the decision to migrate to MongoDB, although we mostly came from more of a SQL world. With that kind of transition, there is always a need to learn a new language and everything else that goes with that. That’s when we came across Studio 3T and we found it very helpful in terms of how to actually learn mongoshell, process it and view the data.”
The Analysis Platform
The SBIA team have built their platform, orchestrated with Apache Airflow, running Python/PyMongo code, importing datasets and querying MongoDB Atlas databases. Studio 3T allows them to tune the platform on the fly by letting them create new optimized aggregations, generate new Python code, and reload it live into the platform infrastructure.
Most used Studio 3T features
Aggregation Editor, Query Code for generating Python scripts, Collections View
Made for scaling up:
- Currently SBIA have 350 aggregation pipelines, querying 1.5TB of data
- An average pipeline has 8 stages, generating 1200 lines of Python code
- The largest pipeline has 21 stages, generating 49,000 lines of Python code altogether
- 6-10 hours a week per team member
- 50 hours a week across the 13 member team