Load cells and big data: opportunity knocks

As business people, big data is nothing new. As part of our everyday work practices, we collect, analyse and create gigabits of data every day without even thinking about it. Big data is simply the data we collect from every available point, as frequently as possible, and use to give a holistic picture of our activities.

Unlike sampling, which relies on data taken at specific intervals or from specific, big data is not selective. Since we can store data so quickly, easily and cheaply, big data provides an ideal opportunity to keep data too, building a resource of information over time.

 

Big data: the numbers

Big data isn’t confined to business either. Data is being accumulated at an astounding rate through processes we all use every day:

• Google processes in excess of 24 petabytes of data every day
• 10 million+ new photos are uploaded to Facebook every hour
• Someone clicks “Like” or leaves a comment on Facebook three billion times per day
• Over an hour of video footage is uploaded to YouTube every second

 

Load cells and big data

So, what has this got to do with load cells? A lot. Load cells enable us to collect masses of data from all kinds of processes, testing, structures, products and human interaction. It is, quite literally, a goldmine of information ready to be tapped into. 

 

Causality and correlation

Businesses may overlook the value of this data because it is ‘messy’. It may contain anomalies, blips, or be chronologically incomplete. However, in the world of big data, this doesn’t matter. The aim is to see emerging patterns, corrolations, and not causality.  As Viktor Mayer-Schönberger and Kenneth Cukier say in their book “Big Data: the essential guide to work, life and learning in the age of insight”: 

“The correlations may not tell us precisely why something is happening, but they alert us that it is happening.”

 

Accuracy v. earnings?

For those of us working to achieve the highest levels of accuracy in testing or monitoring, this can seem like a contradiction. How can data reveal something is happening, but not reveal the why? That’s where the power and earning potential of big data really begins.

 

Posterior profiling

In their book, Mayer-Schönberger and Cukier give an example of the research conducted by Shigeomi Koshimizu, a professor at the Advanced Institute of Industrial Technology in Tokyo. He analysed people’s posture by measuring the pressure exerted on a car seat, sampling the pressure exerted at 360 different point and scoring that pressure out of 256.  The system was so detailed it could distinguish a person just by their seating scores.

The information gathered can be used in various ways:

• as an anti-theft device in cars
• helping improve road safety by detecting shifting positions due to tiredness and altering drivers
• it might also identify a thief by their own unique posture patterning

Of course, in order to detect data to compare with the ‘norm’, sensors are working within the car collecting data, which in turn feeds the big data picture further.

Then the usage becomes wider; could the position of the driver in the seat predict the driver’s state, and be used to warn other drivers in the vicinity that the driver might be sleepy? Many people would find that an intrusion into their privacy, rather than consider it a boon to road safety. 

 

Data usage and privacy

This brings up the issue of data use permission. Nobody needs to ask a machine permission to use data, but humans are another story. Most people involved in clinical research using load cell data will know that their data will be used for the project, and maybe further study reviews. 

What they may not know (and neither do the researchers) is that three years on, any commercially accessible data on their gait patterns could be used to develop better boots for the military, or for designer trainers – or by insurance companies to assess people’s risk of falling in later life. The benefits of the data have spread far beyond the original usage.

As data gatherers, it’s an issue we do need to bear in mind when dealing with data involving human subjects. However, neither should we brush aside the possible uses of our own manufacturing, testing or monitoring data being applicable in a much wider context, for our own benefit and for the wider community. 

So, next time you’re monitoring a process or conducting a test using load cells, keep that data. You just never know when it might come in useful. Our wireless load cell systems enable you to collect, transmit and receive data in real time, ready for later analysis or input into a monitoring system. Call us with your requirements or challenges – we’re always ready to talk all matters load cell.