How many engineers does it take to make a pavlova*? It sounds like the start of a joke – but bear with me.
A recipe is made up of ingredients and method. If the ingredients are the same each time, and the method is followed accurately, the results should always be the same. Change the ingredients (replace kiwifruit with strawberries) or change the quantities and you’ll get a different result: strawberry pavlova, or pavlova for a crowd. Use poor-quality ingredients, and you’ll get a poor-quality result. But, overall, the method is the same each time, and the time it takes is the same too.
If you’re churning out pavlova after pavlova, things are going to get tedious, so you need the best and most efficient tools for the job. Are your engineers and consultants beating the egg whites arduously with a fork – or quickly with a high-powered electric beater? Or could they even design a machine to automate pavlova production altogether?
Compare the consulting work we do at Golder to a pavlova business and there are many similarities. In our case, the data we collect are our ingredients and we follow a recipe to process the data. The machines designed to automatically follow recipes are computer codes, and we can use these to automate the manual steps we use to handle our data.
That’s about as far as this metaphor can run before the cream sours. But it’s a good illustration of the nature of programming and its applications for data-driven projects.
The capabilities and benefits of programming
Across any activity that involves collecting and processing a lot of data, computer programming can automate and speed up tedious, repetitive tasks. The results are often quicker, cheaper, more consistent, more accurate (no more human error), and potentially far more flexible (able to rapidly incorporate different scenarios, criteria or tolerances) and even re-usable.
Ultimately, the larger the volume of data, the more effective it will be to write a computer program to deal with it, and the more you will get out of the data. For the consultant, this frees up time and energy to use our expertise most effectively and creatively, rather than dragging and dropping cells in a spreadsheet. That’s great for the consultant, but more importantly delivers better quality of service and value for money for the client.
When we use the increasing power of technology, we can handle vast amounts of “big data” and accelerated levels of detail in the data. We don’t have to simplify the information to be able to manage it. By maintaining the complexity of the data from collection through delivery to the client, we can communicate much more information.
We can also store datasets that are too big to keep in spreadsheets, pouring them instead into powerful databases. And we can re-use the codes and the data faster and more readily, altering scenarios or tolerances, and picking out trends and meaningful implications.
The tipping point – when to jump
When is the right time to make the leap into the digital future and embrace the power of programming?
It’s a matter of weighing up productivity and time – and this has perhaps been one of the barriers to the uptake of programming. If you’re making a pavlova every hour, you can easily quantify how much progress you’ve made toward your goal. However, if you’re inventing a machine to spit out a mass batch but you haven’t touched the ingredients yet, it’s far harder to demonstrate that real progress is under way, despite the sudden benefit to be gained when the automation comes online.
When you develop a computer program to take on a task, there doesn’t appear to be much progress for a while as you work on the code, but then, in most cases, you achieve instant results once you run the inputs through the completed program. It’s difficult to say with any certainty, however, how long the programming will take. If we can better understand the programming requirements, we can predict if we’ll reach a tipping point and if using a program would be faster in the long run. To accommodate this new approach, the mindset of how consultants track the costs and progress of projects will need to change.
What’s holding the industry back?
It’s easy to see the benefits to be gained from diving into the Internet of Things, machine learning and digital engineering technologies. So why aren’t all engineers and consultants embracing these opportunities faster?
As well as the productivity and costing mindsets we’ve explored above, another barrier could simply be that the capabilities or skill sets haven’t yet permeated through the industry. As more ‘digital natives’ enter the profession, keeping pace with constantly evolving digital methodologies and technologies will become the new normal.
A further challenge, perhaps, is in using our imaginations to see where the technology can be applied. Perhaps it isn’t so much a lack of imagination as the inability to raise one’s head up out of the detail to be able to take a more expansive and strategic view of ways to increase efficiency.
Programming in action
Wherever large volumes of data are being gathered, there will be tasks that will benefit from some level of automation: for example, if you’re gathering hundreds of data points of water depths from piezometers to build a water-table depth profile, or if you’re working with massive amounts of soil data to develop a settlement model, or you’re collecting large volumes of air pollution data. As you continue to collect this data, due to the range of effecting factors, results can change over time. Automation and machine learning allow you to see these results faster and give you more certainty of the quality of the data.
Among many other projects at Golder, we’ve been using geo-programming in our work with seismic hazards, as we run thousands of scenarios to try to predict the timing and probability of future earthquakes and to determine the different ground motion levels that may occur in these circumstances. This involves a lot of data, statistical processing and production of figures. For this work, tasks simply must be automated as it wouldn’t be feasible to do everything manually in the delivery timeframe.
What’s the punchline?
It is hard to say just how many engineers it takes to make a pavlova. But we’re confident that if that pavlova is digital, the answer is ‘fewer’, and the results will get faster and better all the time.
*Pavlova: a classic dessert, named for Russian ballerina Anna Pavlova, consisting of a fruit topping on a meringue base, that is furiously contested by both Australia and New Zealand as a national dish.