As data scientist hiring continues to boom, many organizations report sustained difficulty finding, attracting and retaining data science talent. Even as initiatives to upskill quantitative professionals grow, machine learning literacy remains low in many organizations.

Chief data and analytics officers, or CDAOs, must build development paths that support budding citizen data scientists with the right tools, training and structure. Even organizations that build high volumes of complex and accurate models must diligently foster data literacy and proper adoption of solutions.

Concerted education and culture change are necessary but can be difficult to achieve because of entrenched ways of doing things and the high critical mass of technical expertise required for enterprise data science. Here are steps that CDAOs can take to develop in-house talent and improve data science and machine learning literacy:

Raise machine learning literacy and promote collaboration with data scientists

CDAOs building development paths for data science experts can start by raising the level of discourse around data science and machine learning in the organization. Ensure that all line-of-business leaders and decision makers have a clear understanding of how data scientists create value.

CDAOs must help employees who express an interest or aptitude become familiar with the basics of several machine learning techniques, such as regression, clustering and classification. Data scientists can regularly hold open sessions to discuss projects or aspects of data science they are passionate about. Encourage upskilling individuals to attend regular training, engage in new subjects and enter into healthy competition with peers to maintain their enthusiasm.

Chief data and analytics officers also must raise overall data science and machine learning awareness, adoption and literacy by providing centralized education resources and showcasing existing use cases and success stories, both internal and external. By 2024, 75% of organizations will have established a centralized data and analytics center of excellence to support federated D&A initiatives and prevent enterprise failure.

Foster interconnected data science communities

Talent alignment, career development and talent retention are the primary leadership demands needed to sustain successful upskilling initiatives. Citizen data scientists play an important role for CDAOs when it comes to talent recognition and development. CDAOs must understand citizen data scientists’ persona and recognize the skills that make for good CDS candidates.

Gauge potential candidates’ interest in a data science career and have them complete self-evaluations of their backgrounds to gather an inventory of what skills can be established. Gathering this information will be vital to designing training programs and making technology investments. Typically, the most promising candidates for upskilling have both educational and professional backgrounds in physics, chemistry, biology, actuarial science, computer science, engineering, finance, economics and mathematics.

Design an upskilling roadmap for expert CDS candidates

Foundational data science and machine learning upskilling initiatives must be conducted for CDS candidates. There are three stages to this roadmap:

1. Approach selection and formal training: All CDS candidates must set out a clear vision of what skills they need to acquire and which they need to hone. CDAOs should task their teams to work with human resources to design career paths for citizen data scientists. Formal training for citizen data scientists can be completed in any manner — online or in-person. Leaders must look into which options works best for CDS candidates and develop clear incentives and milestones to measure progress accurately.
2. Experimentation and prototyping: Upon completing their formal training, citizen data scientists can begin experimenting with their new skills and designing prototypes. Leaders must provide an environment for new machine learning practitioners where users can work with desensitized data for trial and error. Place citizen data scientists under the mentorship of data scientists who can review their work and provide feedback. During this stage leaders must ensure CDS candidates begin honing the communication skills necessary for successful data science and evangelize the methodology and potential of their work with machine learning. They should be in regular contact and collaboration with data scientists and sharing their domain expertise with their data science teammates. A reverse knowledge transfer can fill gaps among even the most tenured data scientists.
3. Delivery and operationalization: The final stage of CDS upskilling is the delivery and operationalization of new models. Chief data scientists and machine learning engineers must take an active role in shepherding CDS projects from experimentation to production. During this stage, citizen data scientists should also promote utilization among analytics and exchange ideas for new ones.

There is an abundance of education opportunities and retention challenges that can motivate organizations to upskill their data professionals at all levels and grow their CDS population. It is necessary for CDAOs to build repeatable and sustainable education programs by designing upskilling roadmaps for CDS candidates and expert data scientists. With a large number of tools available to citizen data scientists, CDAOs must navigate this landscape to match diverse users to appropriate solutions and corresponding educational paths.

Peter Krensky is a director analyst on Gartner’s Business Analytics and Data Science team, where he covers data science, machine learning and data and analytics education. He wrote this article for SiliconANGLE. Gartner analysts are providing additional analysis on topics of interest to CDAOs at the Gartner Data & Analytics Summit, taking place March 20-22 in Orlando, Florida.

Image: geralt/Pixabay