Innovations in industrial technology can make an operation more efficient and--potentially--more profitable. This is especially true when it comes to data collected and used both on the factory floor and the administrative office. One of the risks, however, with this new reliance on digital data is over-accumulation of large amounts of unused or inaccessible data in a firm's data lake.
Data lakes are a relatively new way of collecting and organizing data. Data lakes collect raw data and store it in a centralized location, often in a cloud server, which is easy to read and access. Ideally, this would allow users at a firm to readily analyze a set of data in its original form. Compare this type of collection to older forms of organization, like data bases. To be clear, robust data bases are incredibly useful. Firms use them to store and organize their data in a secure location. But they come with a trade-off: they're expensive to maintain and require specialized training and oversight.
Data lakes arose as an alternative to highly specialized forms of data analysis. By employing cloud technology, real-time data collection, and easy access; data lakes are are boons for firms who are looking for ways to increase their bottom lines or who are searching for novel solutions to issues. Unfortunately, novel solutions often come with novel problems.
Tech enthusiasts often deride poorly-managed data lakes as data swamps. A typical data swamp may be spilling over with useless, or at the very least inaccessible, amounts of data. Slimane Allab, SVP of Manufacturing at Cognite, wrote last month in The Manufacturer that, "As much as 90% of all data sitting in data lakes is not used by anyone in the company." There's no citation for this figure, so it may be exaggerated. Still, the prospect of large amounts of unusable data should worry any firm leader, especially considering that making data management easier is one of the problems that data lakes set out to solve.
According to Allab, what data need is contextualization. "Contextualisation makes your data accessible so that everyone in your organisation can draw insights into your data and operations. Assets in your facility may have process variables, work orders, documents, and inspection data across several sources all residing in your data lake."* Putting your data in context can regiment your data into usable formats and access points, removing the free-floating chaos that characterizes data swamps.
Once the data is contextualized, it can be better analyzed and applied to problems. Allab cites one such example in his article about a manufacturer called Aarbakke, which managed to reduce their tool assemblies by 60% and increase their cutting tool efficiency by 10% after properly contextualizing their data pool. Clearly, it's better to have clear lake waters to swim through than murky swamp muck.
Making the Best of your Data Lake
Depending the amount of data in your data lake, even contextualization can become a huge endeavor. Luckily, data lakes, due to their centralized nature, are easily assessed by machine learning systems. Machine learning can quickly analyze, organize, and contextualize large data lakes. It can also offer suggestions based on said analysis and, in some cases, entirely automate problem solving on the factory floor.
In addition, data lakes with good infrastructure are invaluable tools for scalability. As data become organized, siloed sectors can more easily communicate, more users can make use of the data, and novel solutions can be deployed in new situations.
In short, the future of manufacturing is digital. You'll need a reliable advisor in your corner if you're interested in employing a data lake at your firm and successfully deploying its best features. Contact Titan Tech today for a free consultation on how to take your manufacturing operation to the next level.
And we'll see you next week for more tech news.
*The Manufacturer is a European publication. We have kept the writer's spelling intact, as a result.