Data Science Reskilling Challenges: Four Key Considerations

Data Science Reskilling Challenges: Four Key Considerations

This is part of Solutions Review’s Premium Content Series, a collection of contributed columns written by industry experts in maturing software categories. In this submission, Talend Senior Director of Architecture, Strategy, and Delivery Innovation Gareth Shercliff offers key considerations for common data science reskilling challenges.

SR Premium ContentAs the digital world evolves at a rapid pace, organizations must work to accommodate a constant flow of cultural and skills-based adjustments. Rapid technological advancements are impacting every corner of the workplace — from HR to marketing, finance, and beyond — and data is becoming universally prevalent and key to company success.

Data scientists are valuable for their ability to parse through data and craft solutions, but there’s been a critical shortage for years. This makes “citizen analysts” necessary for organizations to properly manage, analyze and leverage their data moving forward.

Citizen analysts already exist in your organization. They need a structured introduction to the basics of data science; a strong leadership-driven approach to data-driven decision-making; empowerment to work in a world of uncertainty; and the desire to understand possible paths forward and then to use the data to point to the best path. This approach will require leaders to reskill their teams in a data-centric direction and make widespread cultural changes that place a premium on organized, healthy data practices.

To establish and nurture this capability across your business requires a combination of repurposing concepts from data science and adopting critical cultural changes to empower and reward citizen analysts for leveraging data as part of their day-to-day activities.

Start by borrowing from the Data Science Toolkit

An individual citizen analyst does not need to be an expert in all areas of data science (and likely never will be), but, they do need a broad and high-level understanding across all skills. That baseline will allow citizen analysts to build more in-depth skills in specific areas as warranted by the problems they need to solve.

Organizations starting their citizen analyst initiative should spend time selecting, introducing and establishing learning plans for the fundamental technical skills for data science, with the most important including data structures, simple algorithms, basic probability, and visualization and machine learning tools. For organizations that already have some data science, or other similar capability, it makes sense to do this in partnership with those existing teams. Establishing a common baseline across the organization gives teams that work with data a common corpus of knowledge and common language around data.

Training in core data science skills is often best delivered through traditional learning plan-based platforms, where the learner is provided with a set of lessons and exercises to practice these technical skills. It is almost always easier and cheaper for organizations to procure training packages for data science, rather than build and deliver their own material.

Beyond data science skills, organizations also need to develop a data-centric culture.

Setting the bar for data-driven decision-making needs to happen at all levels

The perceived value of becoming a citizen analyst hinges on the perceived value of data-driven decision-making within the organization. To firmly establish this value, all leaders need to share how data influences decisions. This especially applies to mid- to senior management, where the majority of the impact of data-driven decisions will be visible.

Sharing data gaps and data failures as well as data successes is important in creating an open culture. This makes storytelling a critical skill. Similarly, being well-versed in the science and art of data visualization allows communicators to share their results in the most highly impactful way, cutting through the usual noise of management reporting and highlighting those aspects of data that are most important. The degree to which the whole organization is successful in becoming data-driven is most impacted by leaders and their communications. Leading by example is the best way to inspire citizen analysts.

Storytelling and data visualization are as much art as science and need to be practiced to be effective. Storytelling in particular requires deliberate alignment with each individual communicator’s personality and style in order to feel authentic. Coaching and mentoring programs work best to develop this skill. This can be done in-house, but more often than not is procured from specialist training providers or through executive education programs offered by many academic institutions.

Working with, not against uncertainty

One of the most fundamental challenges in data science –– but one that is rarely well understood by those outside the domain –– is working with imperfect data. A trained data scientist not only understands that the time and cost of getting near-perfect data is often prohibitive, but also understands the quality threshold required of the data in order to make it valuable to use.

Citizen analysts need to be trained to both interpret data and infer from its characteristics and lineage how useful it will be. Essentially, a citizen analyst needs to develop risk assessment skills on top of understanding the data itself. Citizen analysts who understand the risk profile of their data and the sensitivity of the business to risk and data variations are well poised to leverage that data appropriately.

To master this competency, citizen analysts should be taught a framework for risk assessment, how to gather information on the lineage of the data, and what acceptable cost-benefit analysis thresholds are for the organization. Like with data science fundamentals, if your organization already has a risk management framework, you should align to and train on that. If your organization does not have one, all the larger learning platform providers have courses on risk management, including industry standard ISO 31000. To understand data lineage, citizen analysts need to understand both the concept –– which is covered in most introductory data science curricula –– and be coached to ask for data lineage information. Cost-benefit analysis acumen is best achieved through coaching sessions and internal workshops where participants review scenarios, the data available and then discuss possible decisions they can make based on that.

Without the ability to understand and work with imperfect data, citizen analysts may struggle with over-analysis, or worse still, may not have the confidence to use the data in their decision-making at all, and instead revert to intuition and instinct.

The data reports the news, but what do I do about it?

Understanding the fundamentals and being comfortable with uncertainty are good starting points in data science. However, if citizen analysts still don’t know what to do with what the data shows, they’re still missing the skills they need to do their jobs.

Citizen analysts need to be trained to think through the range of possible decisions before looking at the data. After a range of possibilities is hypothesized, the citizen analyst then interrogates the data to evaluate each option to determine the best path forward. Without practicing and honing this critical skill set, the citizen analyst will struggle to apply the “so what” of analyzing the meaning of the data.

Despite the shortage of data scientists, organizations can gain a competitive advantage by leveraging data through fostering the skills of citizen analysts. Citizen analysts are employees who exist in your organization today who can be upskilled by leveraging self-paced learning platforms for baseline data science fundamentals and risk management framework knowledge combined with coaching and mentoring.

Gareth Shercliff
Latest posts by Gareth Shercliff (see all)