EA models after all are performed with the aim of providing insight into the current operating model or a proposed one.
We’ve been looking at a few examples of how the Enterprise Architecture group can find allies by engaging horizontally with stakeholders – in other words, by finding a commonality of interests, and ways to collaborate, with other initiatives in the organization. In my last post, I described how I’ve seen the Enterprise Architecture group successfully engage and build an alliance with a service quality initiative.
Today I’ll consider a case that I’ve only seen once, but I’m not sure why. I’m referring to the case of data science. To quote Wikipedia, “Data Science is an interdisciplinary field about processes and systems to extract knowledge or insights from data in various forms, either structured or unstructured, which is a continuation of some of the data analysis fields such as statistics, data mining, and predictive analytics, similar to Knowledge Discovery in Databases.”
This description does give us our first clue as to a potential synergy that exists – ‘extract knowledge or insights from data’. If this language sounds familiar, that because it is. EA models after all are performed with the aim of providing insight into the current operating model or a proposed one.
The description actually provides a second clue to possible synergies as well – “either structured or unstructured.” Often, one of the key challenges in analyzing data is understanding the structure of that data, and this provides the other side of the synergy. It turns out that the activity of modeling the current state, capturing the structure of what depends on what, serves to provide a direct input into the analysis that an organization’s data science initiative performs.
At the same, the deeper focus on analytics from the data science team can complement, and to some extent drive, the analysis provided by the architecture team.
So, the efforts involved in modeling an architecture current state (and proposed architecture future state) turns out to have a natural synergy with corporate data science initiatives. Both of them come from similar desires to understand and gain insight, while the slightly different approaches and interests of the two groups complement each other.
This brings me on to the second area of synergy – I've generally found that architects, by nature of the career path that they've taken, can struggle with how to organize and present their conclusions. It's explicitly accepted in every treatment of data science that I've seen, that selecting the correct presentation for the recipient audience is an inherent part of the data science discipline. So a knowledge transfer becomes possible from the data scientists to the enterprise architects.
Almost as if to pay things back, there is a way in which Enterprise Architecture can help the Data Science initiative. Specifically, for data science to be effective data needs to be managed and available, but while data scientists are only peripherally occupied with data governance, it is a natural core function of an architecture group. So this provides a second way that the architecture team can provide valuable assistance to the data scientists - in terms of ensuring that the data that they depend on is managed and governed effectively.
So – as we've seen before, an architecture group can find that they can sometimes find allies in some very unexpected places. As stated before, this kind of co-operation is something I've only seen once – perhaps the buzz around data science is too new – but I'd be surprised if it does not become more common over the next few years.