Keeping Up With Data #89

5 minutes for 5 hours’ worth of reading

Adam Votava
4 min readJul 1, 2022

Data and insight monetisation is a hot topic. More and more companies are realising that there are many external factors influencing their performance — just like weather is impacting demand in the hospitality industry or sales of umbrellas and wellington boots.

It makes sense. It would be a little too confident to think your company’s business performance is independent on anything and anyone else. Sometimes the data you need is (almost) freely available like weather. Sometimes not but there might be someone else who has it.

The companies with in-demand data can (within legal and ethical norms) sell the data to others. The buyer will have to ingest the data, process it, add it into the data flows, build the analytical features, train models, and make decisions they need to make. All of which takes time and carries a lot of operational risk.

Which is why climbing the data product pyramid shown above can be very beneficial to both parties. Buyers will straight away get the answers to their questions, or solutions to their problems. Sellers will get more money. Win-win situation.

That’s why data products are about business, less about technology.

Apart from an article about data products, I’ve included one on the future of AI and one on building the quality from inside out. Enjoy!

  • Yann LeCun has a bold new vision for the future of AI: One of the most influential AI researchers said he hit a wall that forced him to re-think the vision for building human-level AI. He is now proposing the breakthrough might lie in the ability to learn the internal models of how the world works — the ‘world models’. Overall, there are six modules in his vision (world model being one of them and the most complex one). The architecture is very much inspired by human brain — there is a perception module taking in the inputs, cost module predicting the level of discomfort, actor module proposing the actions, short-term memory, and the configurator performing the executive control. Let’s see how far can the AI research community take his ideas, and whether it will have the desired practical impact reasonably soon. (MIT Technology Review)
  • Data Product Pyramid — Implementing a Business Strategy using Data Products: Very well organised article, with many thoughts on data product-related topics. It starts with technology-first definitions of data products — discoverable, reliable, interoperable, [insert the latest buzzword] data. But then it starts bringing data products closer to business: How is it driving P&L? What it tells about the business? What it tells us to do? The data product pyramid (see above) has data in its foundations, but then we have layers of information products (descriptive analytics), knowledge products (diagnostic, predictive analytics), and ultimately decision products (prescriptive analytics) at the very top. The article doesn’t stop there! It talks about how data products are fitting business strategy, how to break down business use-cases into valuable small incremental deliveries, or create a portfolio of business-relevant data products. (Dataception)
  • Building quality from inside out: Adding quality controls is a solution to poor quality, but it doesn’t solve the root cause. Many quality issues result from complex code or complex design. “Less complex code, or even better, less code, means fewer potential issues and higher quality.” I remember an anecdotal quote from my CS class — if there is a bug, look into the part of code you are particularly proud of. The complex design is imho even bigger problem. The whole IT, software, and data industries have built absurdly complex solutions to very simple problems. That’s why we need experienced people for the design. Ideally those that have seen consequences of bad design (typically their own — one learns better from their mistakes). As Petr concludes, we can’t add quality to our solutions, we need to “build it via intentionally designed architecture and better code.” (Petr Janda)

After days of fighting a bug my son brought from the kindergarten (Seriously, what are they playing with? Poisons? Radioactive material? Biohazardous materials?) I’m now looking forward to a sunny weekend and hopefully some cycling and hiking with the family.

In case you missed the last week’s issue of Keeping up with data

Thanks for reading!

Please feel free to share your thoughts or reading tips in the comments.

Follow me on Medium, LinkedIn and Twitter.



Adam Votava

Data scientist | avid cyclist | amateur pianist (I'm sharing my personal opinion and experience, which should not to be considered professional advice)