The art of data science
“Digital experiences can now offer a level of personalisation that has not been possible before. That is what we need for our students exploring courses”
With the launch of IDP Live planned for the second half of 2021, IDP Chief Data Officer Stuart Nickols explains about how he and his team have used data science to build and improve the app for our students.
As my family will attest, I have always been curious and a lover of questions. It is this curiosity that drove me towards data science. I am passionate about digging below the surface to find patterns and logic to seemingly unconnected actions and reactions. I am also fascinated by people and our motivations.
So, it is no surprise I wanted to work in a career that pulled together my desire to play with data with helping people solve their own challenges.
I joined IDP 12 months ago. The company had a clear goal. They wanted to build data science capabilities to ensure students and education institutions could find their perfect match, faster and easier. This was something I could get onboard with.
IDP’s data science ambitions fell neatly into their broader vision of creating a connected community to guide students to achieve their lifelong study and career goals.
With that clear vision in mind, I built a team of some of the brightest and most curious data science experts to help us get started.
One of the biggest pieces of work for the IDP science team has been building the Course Recommendation engine. We have built this function from the ground up – prioritising personalisation and the user experience to ensure its successful uptake for all our students.
Digital experiences can now offer a level of personalisation that has not been possible before. That is what we need for our students exploring courses. IDP has set out to create a curated recommendation engine for every individual user of the app. So, like how YouTube or Netflix serves up content based on where we are and what interests us, IDP can recommend courses to students based on what is best suited to every user.
To create this function, the team had to build a map of higher education data, which built out connections between different courses based on their keywords, content, and student applications and interests.
1. Analyse course outlines
We started with the course content, building an algorithm to read all the synopsises of courses from institutions and linking between the different courses based on similar keywords. This created the outline of the map and the beginning of the roads that linked one course from another.
It is fascinating to see logic of how a person moves from exploring a course on STEM, through to an arts course, for example. It really illustrated how students are interested in courses which bridge different faculties or combine different subjects.
2. Past student data analysis
The team had the main roads and outline for the map, so we then had to add more detail and different routes.
We completed a historical analysis of past IDP student’s data. We have the world’s biggest dataset of student interactions. We attract over 100 million website visits a year, but more importantly, we have a skilled team of experts who have real conversations everyday with students around the world.
This analysis took which courses each student applied to, shortlisted and viewed, creating 1,000,000 more links between the courses. This map was then loaded into the recommendation engine to begin personalising recommendations for users. Think of it like a giant roadmap predicting what the next possible turn a student can make when exploring options.
3. Swiping left or right for personalised support
Now that the engine is up and running, we are continuing to make the engine smarter and more personalised. When a student first creates their profile, they are asked to swipe left or right on 10 courses to help the app understand what type of courses suit the individual student. This helps to build a stronger path for each user, making the course space richer and more helpful for every future user.
Ultimately what we are delivering is trust. Embarking on an international education is huge. It is usually the biggest decision a young person has made in their life. Our goal is to show we are by our students’ side at every step. Right from that first course search match that says “hey, this course might be for you”, through to when they first step into the classroom.
As we start to extend our personalisation, we can extend it to a deeper library of content – produced by students themselves and helping students find calm in the chaos. It is about connecting students into the communities of support out there who will help them get into their ideal course and unlock their full potential.
All of this stems back to curiosity and making sure we ask the right questions.
About the author: Stuart Nickols is leading the build of IDP’s data science capabilities to help students connect with trusted advice, sooner and easier.