On the 10th Day of Christmas, My Boss Gave to Me: 10 Seats of Tableau

On the 10th Day of Christmas, My Boss Gave to Me: 10 Seats of Tableau... 9 AGVs
... 8 One-eyed Wearables
... 
7 Dumb Drones
... 6 Senseless Sensors
... 5 RFID Tags
...
 4 Data Science Boot Camp Graduates
... 3 Used 3D Printers
... 2 Mindless Co-bots
... And Another Vendor Pitching IIoT

Well, today I got yet another holiday gift from the boss, Tableau. Tableau is used in so many places to the point where people that have never used it, know that it must be a good product.

Tableau does not create data (it can calculate new data, but that is a longer story). It doesn’t tame data. It doesn’t cleanse data. What is does really well is take disparate sources of data and visualize them. One of its greatest tricks is that it doesn’t have to move the data. To use a golf term, it “plays the data where it lies.” (Meaning where it is located, not data that lies like a Putin popularity pole.)

So if you step back from this picture, it shows up at the intersections of different systems or data sources where no one data source tells or visualizes the whole story. It can be the canary in coal mine of tech that is cobbled together or it can be the source of groundbreaking insights available to you long before they can be seen in any other system. Generally when you are given 10 seats from your boss though, it probably means you are going to have a fair bit of work to do. Having said that, Tableau is an order of magnitude easier than solutions that came before it. The most important questions you need to be asking are where is the data coming from and what state is it in?

In the world of the 4th Industrial Revolution, it shows up a lot. In general, we are trying to bring together machine and human data in new an interesting ways to reveal new insights. As you see these fantastical new ways to visualize your business, you need to be asking – “hey, can I do that?”  If your data is an ungodly mess, you’ve carved out a lot of work for yourself.

Take the simple example of a factory that has three machines taking temperature. You might want to chart that temperature over time and compare the temperature measured by these three machines. Now lets say that the three machines came from three different vendors. The first thing you’ll need to know is whether or not the data is being collected in the same way. If not, your are going to need to normalize it before Tableau will be able to make sense of it. This is were a lot of heavy lifting comes in. The German machine you bought might be set to Celsius or it might collect two decimal points where as another machine collects only one. The time stamps of each reading could be in seconds, tenths of a second, or milliseconds. Date format for the data when the temperature is collected may be different. If you are in plant leadership, you need to ask, who in my team has these chops?

If your plant is like most, these skills are in short supply. At the HQ they may have a data science team to die for, but trying to get access to them is nearly impossible. This is why 10 seats of Tableau might now be the best gift for a plant. In a previous post in this series, I made reference to the Rime of the Ancient Mariner – "Water, water everywhere and not a drop to drink." In the modern factory, data and insight are like that. You can be drowning in data with no insight you can actually consume.

As you are building your factory of the future, it is important to remember that as hard as data collection can be, it is the easy part. When we walk our customers through how they get to real insight that can propel their business faster, we walk them through the standard charts and graphs we provide as well as how they can leverage our real-time data pipeline. We also help them build custom reports unique to their business.

As an architect of your company’s future, you really have to begin with the end in mind. In our case, if we are helping companies measure the impact of human performance on their results, we begin with the business KPIs that are most impacted by humans. We then help them design their data collection so this data is normalized at the point of collection. This makes charting and graphing in dashboards straightforward, regardless of the analytics tool they use.

Boss, I won’t be returning the seats of Tableau on the 26th. I am going to give it to the data science boot camp graduates you previously gave me. I am then going to see if we can get normalized data from the other gadgets you gave me that we aren’t returning. It still seems like a Lego project, but we are getting there. 

Just two more days of gifts as we wind down to the big day. Can’t wait to see what your last offerings are.