Authored by Dr. Kim Nilsson, Chair & Co-Founder at Pivigo (http://www.pivigo.com), a data science marketplace and AI-as-a-Service company that has delivered over 250 data science projects to 150 organisations globally, as well as Member of the Advisory Board at Corndel.
These days, democracy is very much on everyone’s minds as it is being threatened in countries worldwide. What are the core elements of democracy? Openness? Transparency? Working towards a common goal, following clear and agreed rules? The same values need to be held within corporates in order to motivate their employees, and empower them to deliver great results.
The word democracy is also often used as a slogan for the need to “democratise data” within organisations. There is a growing movement around the need to put data itself, data science tools, and data insights into every employees hands – that it should not be the preserve of a small elite of data scientists. The argument is clear; the more team members aware and knowledgeable about data and its applications, the more ideas generated for how to use data, the more brains applied to the problem, and the greater adoption of data solutions. Simply put, people can do their jobs better.
How do you go about getting the data in front of your team, and getting them stuck in? There are a few considerations, including:
- Technical – how does your team get access, how do they work with the data
- Skills – how do they know what to do and how to use the data
- Governance – how to protect the organisation from any privacy or security concerns
Starting with the first, data is typically either distributed in such a way that it is difficult to get an overview of its existence, or access to it, or it is stored in a cloud database or server that is only accessible by a few technically gifted team members. Furthermore, beyond access, tools are necessary for data manipulation. To solve this problem, many solutions have emerged in the last decade to simplify data analysis for non-experts, such as DataRobot, Dataiku, Alteryx, and many more. These tools facilitate the building of machine-learning applications as well as connections to data sources, and are therefore part of the solution to the democratisation question.
As simplified as these tools are, some knowledge and understanding of the fundamentals of data analysis are still required. Specifically, a good understanding of hypothesis testing, statistical significance, and the risk of bias, as well as an overview of what tools exist in a data scientist’s toolbox and which tool works on which problem. A risk model may require a linear regression, a segmentation problem a clustering algorithm, and a predictive maintenance question may require a time series analysis. How to use which tool, how to tweak the parameters, and how to know when you can trust the outcome are all skills that are required, but also relatively easy to pick up for any analytically and curious minded employee within your organisation.
Of course we must not forgot the question of governance in our efforts to put data and tools in the hand of every employee. A careful review of any privacy or ethical implications of the work you wish to do should always be carried out. It is useful to educate your employees about these considerations so that they may flag anything inappropriate, but responsibility should otherwise lie with a CIO or senior data manager in this case. The most sensitive use cases can then be limited to within a data science team or small group of individuals. Even so, there will be many opportunities to gain value from your data that do not have these implications and employing your “citizen data analysts” will allow a much broader base of talent and brains working on delivering data applications. It will furthermore allow data initiatives to grow from within the team, facilitating the culture change required for going from a ‘data curious’ organisation to a ‘data driven’ one.
Of all of these considerations, what is most critical and what is hardest to get right? In my experience, the hardest challenge – and bottleneck! – in most organisations is skills and talent. Tools are relatively easy to buy and implement, and governance is typically already being taken very seriously by most organisations. Skills, however, tend to be either taken for granted, or the individual is left to take responsibility for their own training. With the pandemic raging, and many employees left temporarily without an active job to go to, this is an excellent opportunity for organisations, and for governments, to support these individuals to upskill and prepare them for the new knowledge economy, including in data skills and literacy.
There are hopes on the horizon for our society, and we have reason to believe that 2021 will be a brighter year for us all. Use the opportunity to make sure it is the year that every employee in your organisation is enabled and empowered to support your digital innovation strategy for the decade ahead!