Home Blogs & News

United Learning Case Study

The below was written by Senna Ma (Project Manager) and Jack Buckingham (Technical Mentor), relating to their time at the Data Science for Social Good.

Project Overview

University enrolment in the UK is higher today than ever before. Despite this, many students across the country do not pursue their aspirations towards higher education (HE), even if they have the academic ability. Schools would like to make sure that pupils make use of their academic potential to achieve their desired pathway, and could help overcome potential hurdles these students face in pursuing higher education – if only they knew which students may need support.

In the summer of 2023, the University of Warwick’s Data Science for Social Good programme (DSSGx UK) partnered with United Learning, one of England’s largest multi-academy trusts, to solve this problem. They worked together to develop a machine learning model that predicts, at an early age, a student’s likelihood of attending Sixth Form. This allows schools to identify students with excellent academic performance but low predicted likelihood of going to Sixth Form as those who might be lacking the confidence to fulfil their academic potential. By identifying students early, schools can provide additional support through supplementary programs to help these pupils. United Learning, who oversee 93 schools with over 60,000 pupils, provided DSSGx UK with a swathe of data. Development and data security was ensured by Aridhia’s Trusted Research Environment (TRE).

Aridhia’s Trusted Research Environment (TRE)

Aridhia’s partnership to DSSGx UK was a key contributor to the success of this project.

Data Access

Aridhia’s TRE allowed both organisations to upload and download data with confidence through the use of the airlock, which gives project leaders control over what data leaves the system. Aridiha’s role-based access control created the trust needed for both parties to work collaboratively.

Development

Aridhia’s TRE enabled DSSGx UK’s global team of data scientists to develop remotely. Using notebooks and virtual machines, data scientists could easily access and analyse the same data resources provided by United Learning.

Aridhia’s support of Gitea (a self-hosted version of GitHub) allowed the team to collaborate the way they are used to. Data scientists had access to familiar tools such as Virtual Studio Code, which further reduced friction, allowing them to put their focus where it matters – on investigations and model development.

Last but not least, a helpful feature was Aridhia’s cloud backup, allowing data scientists to work with confidence knowing that their work is backed up in the cloud, and that saved states are retrievable when desired.

Customer Support

Aridhia’s team provided timely support. From answering basic questions about the interface to specific inquiries about customising the workspace, the team was knowledgeable, kind, and quick to respond.

In addition, Aridhia provides a robust library of documents and videos making it easy to self-serve and learn about specific features.

(You can read the previous DSSGx UK case study blog here.)

Authors

Jack Buckingham joined DSSGx UK in 2023 as a Technical Mentor, providing technical leadership and small group teaching for the team of burgeoning data scientists. Jack has previously worked as both a data scientist and data engineer in an early stage start-up, and is currently pursuing a PhD in the field of Machine Learning at the University of Warwick.

–

Senna Ma joined DSSGx UK in 2023 as a Project Manager. She worked closely with the global team of data scientists to plan and best support the team to ensure a successful completion of this summer’s programme. Having previously worked as a data scientist at a Canadian growth company, Senna is pursuing several passion projects relating to data science and software development.