I think the most important skill for a data architect actually
has little to do with data - and that skill is
systems-thinking. And the second most important skill is
Acceptance: being able to accept and non-judgmentally work with
whatever data asset, system, tool, or technology falls into your
scope of responsibility.
Data-wise, the skill and practice of data mapping is absolutely critical - and it is literally the missing link in data architecture that very few people recognize or are talking about.
I like William's reply and will add that I think it has to do with having knowledge of the capabilities of all of the legitimate data platforms available today and making sure that provisioning is done well. This would mean having some research capability within the company and having the influence to be consulted on for such matters.
Shift away?! RDBMS technology is a bedrock technology! It is the de-facto standard for millions of mission critical applications. Oracle reported a $62 million increase in 2018 for software and cloud services over 2017.
Granted, the shiny new technologies receive the lion's share of
news, webinars, and publicity in the data conferences. The
good data architect should review each of these platforms to see if
they provide value in their company. But don't look for
RDBMS' to fade to black anytime soon.
In Reply to Chris Collins:
Given the apparent shift away from RDBMS...
I agree. Investments in RDBMS remains very high. All other forms of data storage barely register when compared to the RDBMS. The good news is they are all about data and there's room for growth in many data platforms. I actually think the other platforms are underutilized.
Of course, the RDBMS has evolved but at the core of the model is a data page with records (or columns) of strategically placed and accessible data.
This is a timely discussion for me. My son recently asked me exactly what I do as a Data Architect since he is considering a data related degree as an undergrad. So I told him that it depends on whether I am focused on Architecture for Strategy, or Portfolio or Solution Delivery; however, regardless architecture is knowing the right things to do based on business goals, objectives, requirements, risk considerations, and identifying the gaps through transition architecture(s) to achieve the target architecture. It may not be all data related activities either. Change management considerations are very important to realize target state. As a data architect, I am less concerned with the database technology as I view that to be the role of the Data Engineer / Data Modeler, which currently I spend 50% of my time in my current role. My deliverables are more focused on delivering value through the use of data that align / support with business goals.
Really interesting topic.
Quite recently I' m trying to give a response to a similar question in my linkedin blog, but STARTING from a Performance DBA perspective
Paolo Filippi[All DATAVERSITY Members] @ Apr 13, 2019 - 11:55 AM (America/Pacific)
I think that it means a lot more than just the technical thought
leadership and control that we traditionally think of as being the
core of what an architect does. Regardless of how much we're
working with RDBMS vs. NoSQL databases, our jobs are increasingly
about the "soft stuff." I'm starting to think of our role as being
something of a type of change manager with specialized systems
Aaron Fuller, CBIP
Founder & Principal Consultant
Superior Data Strategies
Jose Mari Taleno
As for me, being a new Data Architect, it means being able to come up with a sound Data Strategy for the Business Intelligence Team of the company. It means heading the design of the Data Architecture considering several sub components like quality, security and scalability. And most of the time, still getting the hands dirty in orchestrating the ETL to have a data model for the Data Warehouse in which should be the main source of facts of reports and dashboards with consideration on performance and user experience. Am I still on the right concept of a Data Architect?
I totally agree with Stacey on this. For me Data Architect role is less about technology and more about providing value to business based on their requirements. Also, data architecture should include the whole DataOps paradigm not just the Data Management.
A Data Architect strategically positions systems, data, data stores, interfaces, and controls to work in concert to achieve business data objectives including identifying data to be protected and governance according to company policy.
I think that data architecture does do strategic implementation. It seems that there is a managerial component. I would check out this article: https://www.dataversity.net/data-architect-vs-data-engineer/.
The role of a Data Architect is still evolving.
This is from LinkedIn..
A skill every data architect should master - "The person who can understand data, explain it to the business and make it actionable will become the most valuable employee of the future"
credits: Scott Taylor, Michael Dell, Marco Aurelio Cavalcante Ribeiro
That's a classic definition of and argument for a Data Scientist isn't it (or"why you should want to be one")? Are they the same role as a Data Architect? In my view there is a direct relationship but different perspective or focus: Architects plan and design the data strategy and infrastructure to support those who need to understand data and use it innovatively to create value for a business, like Data Scientists for example.
I am reading Jikku's comment that understanding data, explaining business and making it actionable are important skills for a Data Architect. It also is said about data scientists. I would argue that skill is applicable to both. Yes, architects plan and design the data infrastructure. At the same time, they need to check that this design is something the business can use and explain the larger picture to them.
We can make a separate topic on differences between a data architect and data scientist. Would that work?
Read “Data and Due Process: When Algorithms Go Awry" by Michael Shaw.
Data Architecture is someone that helps move the project along. It is someone on your side. Collaborative. Not a “no” person.
Data is neutral. What is not neutral is the interpretation of that data.
We should be conscious of confirmation bias in that we don’t want to fool ourselves into thinking we are right when our initial emotions may need to lead us to more analysis.
We need to check and recheck because data depends on context.
Thank you Lawrence. I like this. I read Michael Shaw's article and agree that data depends on context.
[login to unmask email]
Freelance Production Assistant
Freelance Data, Technology and Science Writer
Michelle Knight[All DATAVERSITY Members] @ Dec 23, 2019 - 08:39 AM (America/Pacific)