Choosing your next read can be an agonizing process. With limited time and infinite choice, readers regularly cite the difficulty of selecting the “right” book as a barrier to reading more.

To help make discovery easier, Rakuten Kobo is harnessing the power of big data on their eReading platform. With a catalog of more than 5 million books and millions of readers around the globe, Kobo is able to take voluminous amounts of data and process it into helpful recommendations.

“Modern recommendation systems, no matter how technically sophisticated, are based on a very simple foundational idea: similar people like similar things,” explains Kobo’s Director of Big Data, Darius Braziunas. “If you see a bookshelf and recognize some of the authors and books on it, it is likely you will enjoy reading the other books as well.”

Kobo analyzes the titles in a user’s library, and automatically extracts the main topics of interest – using unsupervised machine learning techniques – and creates a weighting of these interests based on their relative importance to the user. A user might be interested in both science fiction and biographies, but if the last few books she has read are science fiction, the recommendation system will assign a higher value to that topic, leading the user to receive more sci-fi recommendations than biography.

Big data is enabling Kobo to utilize machine learning to improve personalized recommendations for users.

Personalized recommendations take into account multiple data points – not just recent purchases.

So-called “machine learning,” the science of programming computers to learn by themselves, allows Braziunas and his team to offer recommendations to readers based on the most relevant real-time data. Thus, recommendations a customer sees may be very different from one day to the next (in some cases, even hour to hour).

“Our goal is to make the entire website personalized, to customize to individual interests, whatever those may be.… To do this we take into account a variety of data points – offering what we hope is a richer, more relevant list of suggested reads,” said Braziunas.

Unlike other e-commerce sites or bricks-and-mortar bookstores, Kobo looks at more than just purchases when generating recommendations, making use of actual eBook content, metadata (author, series, genre), and other attributes such as popularity, ratings and publication date.

Kobo looks at more than just purchases when generating recommendations, making use of actual eBook content, metadata (author, series, genre), and other attributes (popularity, ratings, publish date, etc.).Recommendations are delivered via several channels – the homepage, product pages and emails.

These recommendations are then delivered to users via a variety of channels across the platform. There are personalized lists that live on the homepage, “Next Read” emails sent when a user is close to finishing their current book, related books (“people who read this also enjoyed…”) on each product page and a host of others.

“Ultimately, machine learning algorithms are a tool,” Braziunas explained. “We want to help people find books that they will enjoy – and big data is an extremely effective way of doing that.”