To make the most effective use of your data science team, employ the skills of a professional editor.
Editors for leadership and management
Just like you have experts in leadership, management, and of course analytics on your data science team; you must have experts in communication. Editors articulate thoughts and ideas into effective communication media, and there's no strategy or data science team that doesn't need this. So, without a specific role to play this part, editing often becomes a group exercise among your most expensive and least appropriate resources. I've often seen managers unofficially anointed to this thankless role, and even if they are good, wouldn't you rather have them managing the team? Worse yet, the manager's draft now goes up to the leaders who must now spend their time in edits instead of setting the course. This is a gross misuse of valuable resources. With relatively little additional expense, hire an editor to worry about content - they're the experts.
Perhaps the most obvious and direct application of editors is in the leadership function. Leaders are responsible for articulating the vision and helping the organization through the anxiety of unchartered waters. Having a big data analytic component to your corporate strategy makes this task even more daunting. Although many are coming around to the idea of embracing analytics in a corporate strategy, it's still a black box to the masses. A good editor can help explain unfamiliar concepts like machine learning and natural language processing in a way that makes the rest of the organization relate them to the company's future.
Although leaders obviously benefit from having an editor around, management could really use a hand as well. Managers are busy, busy people - busier than anyone else on the team. The last thing you want to do is tax them with another responsibility; however, it seems like managers are the de facto editors when nobody else is assigned. Communication is vitally important to managers - it's they're primary tool for harnessing all the complexity of a strategy, especially one that involves big data analytics. A manager should be able to throw rough ideas and bullet points to an editor, and in return, get clean and clear reports and presentations that can be shared with leadership, the data science team, and the rest of the organization. The trick here is to make sure your editor has enough talent and context to follow what's going on. If the editor's output needs to be edited by the manager - that defeats the purpose!
Editors and data scientistsThere's a special kind of editor that can add a great deal of value to the data science team itself. The role of this editor is to clarify the work produced by the data science team. For instance, your data science team may be tasked with qualitative analysis against your operational data. One of the key outputs of this exercise is a research paper that explains their findings. The editor in this case must ensure this paper is well written. Although the goal for the editor is to make this type of work generally consumable, the main audience for this editing is other data scientists.
Data science editing is not confined to research papers. Remember, data scientists are not only writing papers, but they're also writing code. Having a deep background in computer science, and having worked closely with other computer programmers, I can assure you that programmers have a bad habit of writing unreadable code. I shamefully admit that, on occasion, I've opened up my own code after months or years, only to find a foreign language staring back at me - even though I originally wrote it.
One method of overcoming spaghetti-code syndrome, that's popular in the agile community, is refactoring. Refactoring is an exercise where programmers rewrite functioning code, for better readability and design. That sounds a lot like editing to me! Why not have an editor (or team of editors) do nothing but refactor data science code? That way, the data scientists can continue cranking out code without worrying about unreadable code or slowing down to refactor.
At first glance, you would think that the only people that could do this type of editing are other data scientists, but that's not entirely true. Sure, you need someone with special skills; however, it's easier to edit a book than it is to write a book. Your data science editors should have excellent programming skills, and excellent communication skills, but their analytic skills don't need to be superior. As long as they can understand what the data scientists are trying to accomplish, they'll be fine. The key to making this work is a good testing infrastructure. This prevents the disaster of editors, with good intentions, breaking functional code. By the way, this doesn't only apply to code. On more than one occasion, I've had editors improve my writing, only to mess up the message. It happens. That's why you should always have a final review with the authoring team. However, if your editor is good, it should be a quick and painless review.
If you care about communication, and its role in your big data analytic strategy - which you should - then hire communication experts to do the job. Although you could tax your leaders, managers, and data scientists with their own editing; it's much more efficient to have editors on board. Like any other expert, they're not only better at producing high-quality communication media, but they're also quicker and cheaper. Editors can help out with leadership, management, and internal data science communication; however, it takes special skills to work on a data science team. Make sure they know enough to follow along, translate difficult concepts and ideas into understandable media, and write code if necessary. Take some time to think about who's doing all the editing on your data science team right now. If it's your leaders, managers, or your high-powered data scientists, you might want to edit that practice out.