Data mining word cloud

+ Digitising historic data to analyse it using today’s advanced techniques could reveal new insights.

There are many great examples of firms like Arup getting value from big data – large data sets characterised by their velocity, volume, and variety. These often involve digitalising traditional workflows to sense, compile and analyse new data. What about old data that already exists?  How can we understand whether it’s worth digitising historic engineering data?

One of my favourite examples showing the value of digitising historic data is the story told in the book and film Moneyball. This explains how the Oakland Athletics baseball team analysed over 150 years of baseball statistics to field a team that could better compete against richer competitors. They reasoned that an evidence-based approach, dubbed sabremetrics, was better than a hunch when it came to picking players.

So where might similar value lie in engineering? Let’s take geotechnical engineering as an example. Over more than 50 years Arup has worked on thousands of building and infrastructure projects around the world where ground investigation has led to ground models to support foundation design. Most of the investigation data and models are on paper or in scanned archives, not in a digital format that can be readily integrated into database technologies for use today.

Costly to create at the time, the ground models still have inherent value today. Getting all of this data into one model is something that has been talked about since I started working in this industry 20 years ago. But the cost is prohibitive and the value (as we can envision it now) isn’t always clear. We often get bogged down in the details and now, 20 years later, all of that data is still languishing in different formats and archives.

Mining company Rio Tinto has digitalised its traditional workflow analysing new data from equipment sensors to improve productivity, and it has also sent all of its historic exploration investigation data to its big data centre in India for digitisation. These have discovered targets missed in the past. The result is that Rio Tinto is developing new and enhancing existing resource models – generating a return on its investment.

As BIM is becoming a digital expression of our engineering design, there is an ongoing conversation about what to do and what value it will bring. I see a convergence between big data and BIM, but that’s new data – we still have valuable historic data buried in our archives.

We need to bite the bullet and digitise everything, send our archives to a cost-effective processing centre and trust that data will return our investment by telling us things we can’t envision now.