Penguin Random House UK: The Power of Meta Data

Written By Rupal Sumaria is Head of Data Governance and Toba Shittu is Data Governance Analyst at Penguin Random House UK.


We both work as part of the Data Team at Penguin Random House UK and are passionate about the ways we can gather and interpret data.

Recently, we teamed up with the talented team at UKBlackTech to deliver an online event The Power of Metadata on 17th June. Our goals were: to explain the capability of metadata, discuss how it can help improve diversity and representation; and hopefully inspire some of our attendees to consider a career in data science.

Here are our top five takeaways from The Power of Metadata if you missed the event and want to watch it back – follow this link

Metadata is a vast universe

Metadata describes and provides information about other data that makes it findable and easier to work with. It can be hard to explain the brilliance and complexity of metadata, which is why one of our speakers shared visuals of this model: 

This was presented by Kit Thwaite, Head of Data Science at Penguin Random House UK. It may look like stars in the universe, but in fact each dot represents a single book in Penguin Random House UK’s catalogue.

This model shows what strong metadata can provide: clusters of similar themed books drawn together by common threads such as authors or genre – the closer the dots are, the more similar the book. Data like this helps us to make personal and unique book recommendations for a reader based on their preferences. It’s not just about what people want but how they feel, and as Kit says, for us at Penguin Random House UK, an important part of what we do is listening, tapping into the wider conversation, and adapting metadata around what we hear.

We can also use this visualisation to imagine what poor metadata could result in: a vast array of books with missing character traits. Kit likened this to a metaphorical warehouse with a pit of books piled high with little discernible identifiers, making it impossible for a reader to find what they’re looking for

Metadata is the biggest “little” problem in the music industry

Joshua Babatunde (Co-Founder of Creditz Music) gave a fascinating talk about his start-up Creditz Music and how inconsistent metadata is resulting in missed credit for musicians. Currently data for different music tracks is held in hundreds of different places in different formats – any mis-spelt or missing data in one database could result in a musician not receiving the credit they deserve or even missing out on royalties. Creditz aims to change this by creating one central database, which would be a game changer for the industry.

Data by itself is meaningless

“Data’s not new… we’ve been collecting data for as long as humans have existed.” Dr Suze Kundu.

Nowadays, we can appreciate the insight that we can gain from data, but it still needs people who understand the rules of the data to interpret trends and findings and keep data simple and relevant. 

Dr Suze Kundu, Digital Science, referred to the idea of data citizenship, where it’s not just about what you get from data, but also about what you give back. She talked about how Metadata should always be FAIRFindable, Accessible, Inter-operable, and Reusable

Metadata could help improve diversity and representation

Marco Lau, Director of Data Science and Analytics at Penguin Random House UK, discussed how it’s very easy to measure commercial success via metadata, but it can also be a tool to help answer questions around how our books represent all voices in society. How do we create books for everyone?

All the speakers shared in the sentiment of needing better regulation, ethical guidance, and transparency in the use of data. At Penguin Random House UK, Marco strongly champions his data science team to become more ambitious and transparent in what they create to ensure it is equitable and ethical, particularly in pricing or recommending books.

Now is a good opportunity for all companies to start thinking about cleaning their data and analysing their processes to make sure that their algo-rhythms and processes aren’t perpetuating biases.

Andreas Arnold, CIO at Penguin Random House UK, believes that ultimately, accuracy and quality of data builds trust and trust is what you need to make decisions, and to ensure the outputs of your data are FAIR.

It’s about passion, not experience when it comes to starting out in the data science field

  • Passion
  • Curiosity
  • An analytical approach
  • A willingness to learn

Our panellists agreed that having the above traits is more important than having a statistics-based or data science background.

“Whatever really excites you, there will be a data science and analysis job within that.” Dr Suze Kundu

It’s incredibly powerful to work in metadata…do not underestimate how much you can change getting into this field.” Andreas Arnold.

Final thoughts

A big thanks to all our panellists/speakers for sharing their insights, and to Mark Martin for hosting. It was an inspiring event, and we hope everyone enjoyed it as much as we did.