Marketing + Data Science, Business + Data Science: natural duos that are naturally integrated. What about Engineering and Data Science? You can almost guarantee that every engineer will consistently come into contact with data, no matter the engineer’s focus. Data will always drive their decisions. And often times, the data available to engineers is expansive; yet many engineers are unequipped to handle the data.
Engineers make data-driven decisions daily. It’s vitally important that the average engineer is sufficiently competent at gathering good data and properly interpreting its meaning. But most engineers aren’t equipped to handle data at scale. Most engineers don’t have the background or knowledge necessary to scale data pipelines or automate their efforts. This often times results in mass data gathering but data rotting away with little to no use. The data will only be drug out when something goes wrong with a piece of equipment or batch in a process. Either engineer need to step up their game or data scientists and data engineers need to find their way into these teams.
Plant level engineering is one area data scientists are rarely found. In marketing and business functions you’ll find data scientists frequently in big companies. But for whatever reason, plant engineering rarely includes engineers, statisticians, or even computer scientists capable of building data pipelines and analysing data at scale. This is shocking considering the sheer amount of data gathered by sensors at the plant level. It seems that in many places the only use for data from these sensors are triggering alerts when the process values (PVs) exceed some threshold.
Increasingly, companies are paying others to come in and engineer solutions for automating their data gathering and analytics. But is spoon-feeding engineers’ answers out of thin air solving any problems? Are these answers that relevant in a constantly changing environment? And if not, is Company A prepared to bring in Company B every time a new model needs built or an old model needs to be refined? Sounds extremely pricey.
First, hire data scientists and data engineers alongside engineers in manufacturing plants and design engineering firms. Leverage the statistics and programming backgrounds of these data scientists alongside the process knowledge of the engineers and consistently implement solutions that are taking advantage of available data. In a plant with let’s say 25 engineers constantly responding to and designing new solutions, 2 data-savvy individuals would go a long way to driving decisions in a more data conscious fashion.
Second, train your engineers to be more capable of the basic skills needed to engineer some simple solutions. Most engineers have significant exposure to statistics and math through various coursework. Most engineering curriculum cover at a minimum: basic probability and statistics, distributions, confidence intervals, sampling, correlation, and regression. And the average engineer shouldn’t have too difficult of a time picking up some basic programming in Python. Engineers are also perfect candidates for data science.