It's very likely that the company you work for is looking to hire a data scientist. I am sure the search is difficult if not impossible. Demand for this new (and yet surprisingly well-established) role has pushed salaries up beyond many budgets. Those with the presumed qualifications have the luxury of choice.
But if you lack the budget or the cool factor to attract data scientists today, hold your disappointment. There are sound and practical approaches to machine learning which build on your teams' current skills. In particular by advancing existing data literacy to new levels you can develop what some have called Citizen Data Science as a practice.
There are three key elements to this pragmatic strategy ...
Data literacy is a team sport
Looking at traditional literacy, we don't acquire the skills of reading, writing and arithmetic by ourselves or for ourselves alone. We largely read what others have written, and write for others to read. In this way, literacy and numeracy reshape societies quite as much as the lives of individuals. We may complain about filling out government forms, or find ourselves bored reading and writing memos at work, but a literate society functions better.
Organizations with higher levels of data literacy function better too, and not only because some smart individuals (the data scientists perhaps) have their own unique insights.
A data literate sales team, equipped with good Business Intelligence, knows that "average deal size" is not a good reflection of the business when a few very large enterprise contracts skew the numbers. They are in a better place to act on strategy without missing the big picture. A supply chain team which understands the ambiguities of sales projections can be more agile and more cost effective in supporting production.
This matters, because as Alan Webber, the founder of Fast Company magazine, wrote, back in 1993: Conversations are the way workers discover what they know, share it with their colleagues, and in the process create new knowledge for the organization. In the new economy, conversations are the most important form of work.
In an information economy where market insight, price comparisons and competitive intelligence are widely available, data literate conversations can make a critical difference.
Data Science is still Decision Support
The business technology industry appears to love definitions. I read numerous taxonomies spelling out in devilish detail the difference between, say, Business Intelligence and Data Mining and OLAP.
Years ago, we had an excellent term for this critical practice: Decision Support. I still like that label, because it tells you clearly what the system does: we support you in making better decisions.
Data Science, which has over the years grown out of statistics and operations research and other analytic practices, is still, at its heart, Decision Support. This matters, because it makes clear that there is no sharp dividing line between Data Science as a practice and other analytic techniques.
If you are comfortable with averages, means and medians and you're getting familiar with standard deviations - which many people know from high school or business school - then you're closer to Data Science practices than you might think. You will certainly be able to take part in data literate conversations. And you will bring to those conversations essential business knowledge which a newly hired data scientist might take months or years to grasp, let alone codify.
But let's have a quick reality check. You are not going to be competing with the latest artificial intelligence breakthroughs at Google at any time soon. Amazon's powerful algorithms are safe from your competitive insights for now. But most data science, in most companies who are taking it seriously, is not so advanced. Mostly data scientists are making useful, but incremental, improvements to decision making. And so can your teams, even while walking before they learn to run.
Smart teams grow around smart tools
The final critical element which teams need to develop literate data science from their existing resources is a good toolset. There are two useful approaches which can really help.
The first, is to use preconfigured vertical or horizontal models. For example, Big Squid's MLKits provide models for numerous business functions which can get you started with an efficient workflow by mapping your own data sets to their recommendations. A data literate team can have easily understood models up and running in days which might take weeks to develop from scratch.
A second approach is to integrate Automated Machine Learning with your existing Business Intelligence tools.
As you develop your dashboards and reports with tools such as Qlik Sense, you are in effect, preparing a clean and accurate data set which is excellent source material for machine learning. However, you still need to build and test models for predictive and prescriptive analytics - and surely that is where you need a data scientist? In fact, automated machine learning can help considerably. Big Squid's Kraken platform can choose appropriate algorithms, set the best parameters, and score the outcomes for a business process that you easily define.
Most importantly, both of these approaches are supported by Kraken’s collaborative features - remember this is a team sport! - and the ability to easily put models into production, supporting and automating decisions alongside your BI platform.
Your team today and tomorrow
So, do you need to hire a data scientist? Perhaps not yet. But think of it this way. Even if you hired a data scientist right now, could you support them without the skills, practices and tools we have already described? In too many cases data scientists end up isolated and frustrated: struggling to make their work understood and to get into production in companies which have not developed the data literacy needed to engage with them effectively. Working with a knowledgeable team like Big Squid can help to drive more data-centric conversations into your organization and increase the literacy of your team.
Growing your data literacy, by extending your existing BI practices with platforms like Big Squid's Kraken, enables you to grow a community of practice today. That's good news in itself. Even better, come tomorrow, when you are ready to hire a more specialized data scientist, the market will be ready with all those eager students currently filling the STEM courses at universities, and as demand meets supply, the relative costs of employing them will even out too.
Today is a great time to start building your data science literacy. Tomorrow, your business will be thankful for your foresight.