It seems like an innocent enough question, however, when I hear this question it always triggers an internal debate that starts with one of: Do they really want to know?, or How much time do I have?
I am a Data Architect, but I often just say that I work in Information Technology. Sometimes that’s enough if the question was just a pleasantry, but once in a while I hear the response, “What do you do in I.T.?”, followed by, “OK… What does a Data Architect do?”. You would think that after more than a decade I would have this, but I always find myself somewhat at a loss for words at this point, because the answer is either fairly involved, or somewhat trite.
Data Architects architect data. While accurate, this statement doesn’t say anything. Of course we architect data, but what does that mean, in terms that someone who is not a data practitioner can understand.
Like any architect, a Data Architect designs things. The problem is that everything I design is virtual. It exists in some incarnation of, for lack of a better word, cyberspace. Analogies then. Analogies are one of my favourite tools, but to work they need to be recognizable, and they usually break down at some point, so you need to recognize how far you can take them.
Analogy
The analogy conversation usually goes something like this…
“It’s like a building architect, but with data”
To which I hear a response like, “So… You design data then?”. Followed by, “I don’t get it.”
“OK, so more like, A data architect is to data what a building architect is to people.”
“But a building architect doesn’t design people. Nope. Still don’t get it.”
I pause, looking off into space. “I design structures to hold data in the same way that a building architect designs structures to hold people.”
“OK…?”
“A building architect doesn’t just design buildings to hold people. They design buildings to hold people for a purpose or function. That’s why a house has a very different design than an office tower, or a hotel, or a factory, or a jail. They all hold people, but the purposes are very different. Also remember that the same people may, at different times, inhabit different buildings. The people, plus the purpose is what guides the building architect. Through in a bit of artistry, and you end up with the design for a particular building.
“It’s like that for data too. Data can be used for different purposes. Like when you buy something at a store, they collect data on what you purchased, how much you paid, and possibly who you are. This data initially has the purpose of producing a sales receipt. But it’s later used to adjust inventory, and aggregated to create sales graphs, and run through calculations to predict what to order, when, and how much. Even though it’s the same data, there are different structures that best suit each of the different purposes.”
This is about the time that they say something like, “Yeah… [pause]… My brother-in-law is in IT. He’s a developer”.
Back to the drawing board
Building Architects have it easy. They can point and say, I designed that building.
So I need to take another tack on this. What does a Data Architect do? What do I do? Like a building architect, I do design structures for a particular purpose, but to hold data. But that’s just part of the job (the data modeller role). Unlike a building architect, my goal is to help people understand data. More specifically, to understand their data; how it flows; how one bit of data relates to another bit of data; what they can do with their data. Once people start to understand their data, they start to realize that, far from being a byproduct of doing business, data can be (and should be) viewed as a very valuable asset to their enterprise.
Once we reach this level of understanding (here comes another analogy), a data architect can start to tune the data. This is similar to an engineer designing a vehicle. The design of the vehicle is tuned to it’s purpose. A tractor trailer has very different design considerations than a family car, or a race car, even though they may all be driven by the same person. The structures are different because the purpose is different.
Data is tuned for particular uses by moving it into different data structures. The structure used for a point of sale system is very different than the structure used for guiding decisions on what new product line to carry next year.
A New Definition
What does a Data Architect do?
A data architect facilitates a company’s understanding, and development of data resources. They then work co-operatively with the company to tune the data structures and content to optimize the data’s value in achieving the company’s goals.
