Democratizing Machine Learning at McGraw Hill Education

[DanielleR]

Democratizing Machine Learning at McGraw Hill Education

Name: Fabio Italiano

Title: VP Analytics and Information Management

Company: McGraw Hill Education

Overview of Use Case

Learn how McGraw Hill leveraged machine learning and advanced analytics techniques in Alteryx to empower business users with advanced R packages without having to write a single line of code. Hear how they applied teachings from the Udacity Business Analyst Nanodegree course created by Alteryx to quickly build a community of "business data scientists" that are actively engaged and invested in the success of a governed analytics roadmap.

Describe the business challenge or problem you needed to solve

Over the past five years, McGraw Hill has completed a digital transformation. Today, most higher education revenue is in the form of digital offerings, and the Analytics Office required a way to provide sales and leadership actionable insights to contribute to margin and revenue improvement. The team wanted to empower business users with dashboards and insights that they could leverage to make impactful business decisions.

What is the Analytics Office?

The Analytics Office at McGraw Hill aligns with the business by providing actionable insights and reporting that contribute to margin and revenue improvement. These reports provide a framework to monitor the business with leadership and operational dashboards. The Analytics Office struggled with scarcity of resources and reporting thousands of rows of data within Excel. This was simply not an efficient process. The team wanted to be able to leverage machine learning to impact these reports and dashboards.

What is Data Governance?

McGraw Hill was looking to establish more governance for their data. They wanted to have a methodology in place to collectively identify and improve their data models. The focus of their data governance process is to increase data quality, data uniqueness and data accuracy. The team wanted to stop guessing and save time.

Describe your working solution

McGraw Hill was introduced to Alteryx at 2017 Inspire and quickly joined the Nano Degree by Udacity Program. This program provides initial Alteryx training through our partnership with Udacity. McGraw Hill is currently using advanced machine learning applied to learning platforms. The Analytics Office at McGraw Hill has established a machine learning platform with an intricate algorithm and embedded AI. 

The Analytics Office is leveraging Machine Learning in Alteryx to provide timely updates to sales and leaderships teams tasked with moving the revenue needle. The Analytics Office runs three primary workflows in Alteryx. The first is to run model validation with comparison. The second is model validation on the sample. The third is what goes into production.

Describe the benefits you have achieved

The team gives an update to program sponsors monthly.  Each quarter the data science team walks through the workflow to provide data and insights for sales and leadership. The data scientists would have to write over 50 lines of code to do what they are doing in Alteryx.

What is Machine Learning?

In machine learning, the system learns characteristics from data in almost the same way a human would learn facts if speaking with a group of individuals.

 - Machine learning has been around for long time, it’s the way academic research works: First you test hypothesis on a set of data, then validate the accuracy of the models on a sample data set, after which you apply the models / findings to predict the future.

- Machine learning allows you to process more data, leverage simpler programming languages (R, Python) or use tools that don't require any coding at all, such as Alteryx!

- We are better served explaining the basic machine learning concept with an example: we have a room full of people and we would like to build a model to predict who will watch the football (soccer) World Cup this summer or who is married or who has children based on the information we collect.

The team is using machine learning for advanced analytics such as predicting customer churn. Additionally, machine learning is used to better understand what features of a learning path are driving McGraw Hill revenue. Certain paths, instructors, or content pieces are flagged at risk when they fail to produce revenue. From there, the team can better optimize offerings for their customers by understanding what variables are impacting sales.

Related Resources

The entire PowerPoint presentation can be found attached below. Additionally, visit this link to watch the entire recorded session.