Are you ready to know what the future holds for data science and analytics? As we enter 2023, the advancements in technology and the growing importance of data modernization and data security are set to revolutionize the way data professionals work. In this blog, we’ll take a look at our top 10 predictions for data science and analytics in 2023 and explore the implications for businesses and organizations looking to stay ahead of the curve.
1. INVESTMENT IN DATA MODERNIZATION TO IMPROVE EFFICIENCIES
Investing in data modernization is crucial for organizations that want to stay competitive and successful in today’s data-driven world. We predict that many organizations will continue to increase investments in data even in the wake of potential economic uncertainty, reflecting the continued perception that data is an essential business asset to organizations. From our conversations with the Burtch Works network, including executives within data science and analytics, it seems the main target of data modernization begins with initiatives to transfer data from outdated or siloed legacy environments to cloud-based environments. And while cloud migration is nothing new, we anticipate additional resources toward data observability capabilities to enable real time and more efficient detection of data cleaning, governance, and transformations.
2. PRIORITIZATION OF DATA SCIENCE END-TO-END SKILLS
Data science leaders will begin to prioritize soft skills like strategy, data storytelling, and framing problems, with greater reliance on technical capabilities from applications, systems, and programs. As companies strive to become data science savvy, with attempts to more fully embrace the move from Business Intelligence to AI, we expect data science hiring managers to demand stronger non-technical skills, including communication, the ability to engage stakeholders, and drive change via upward influence throughout an organization. Framing a business problem before models are built will become an increasingly in-demand skill. Completing full cycles of data science only to produce a solution which does not accurately, completely, or truly address the business problem in an actionable manner will need to dramatically decrease to keep data science and analytics capabilities positioned as valued assets and a competitive advantage. Other less technical skills, including data science strategy, will more fully enable end-to-end data science.
3. DATA SCIENCE STORYTELLING BECOMES MORE RELEVANT, ACTION BASED, AND TARGETED TO THE APPLICABLE AUDIENCES
The ways we own the data science storytelling will transition to more personalized, action-based, and relevant content and delivery. Much like the gravitational pull of models to support personalization within retailer, CPG, healthcare, and other application areas, communicating data science solutions will need a more streamlined, personalized, and relevant story to stimulate the acceptance and implementation phase. Guidelines, texts, and other sources, such as Bill Franks’ book, “Owning the Room,” will become more a part of necessary training to ensure that data science professionals are well-equipped.
4. ACADEMIC INSTITUTIONS TO VALUE AND DELIVER EXPERIENTIAL LEARNING OPPORTUNITIES VIA CORPORATE RELATIONSHIPS
Academic institutions will put more focus on experiential learning and engagement with the corporate and practitioner community. We’re already seeing many major universities develop data science centers or institutes, which are intended to serve as a platform and connective tissue between academic stakeholders and the practitioner community. Engaging the parties on issues such as academic program development and advisory assistance, providing thorough leadership from academia to industry and vice-versa, as well as other benefits to each party will encourage a more connected and collaborative network of academics and practitioner groups.
5. DATA SCIENCE SOLUTIONS WILL CONTINUE TO TRANSITION AS A WORK PRODUCT
Data science as a product will increase and continue to develop the notion of an operationalizable solution, with the ability to build these solutions and manage them at scale. The former practice of building completely customized, from scratch, once-and-done models will need to decrease to maintain more of a operationalization mindset. Treating the solution more like a product helps accomplish these goals, and this may mean an increase in development of SCRUM teams for building data science products, and could result in updated terminology, skills needed related to product development, and training for the life cycle of a data science product.
6. SENIOR LEADERS WILL TAKE STEPS TO ENSURE MORE ETHICAL TREATMENT OF DATA AND MODEL DEVELOPMENT
It is likely that ethical standards, boards, and legal guidelines will continue to evolve and become increasingly stringent as the use of models and solutions becomes more widespread and impactful. As such, organizations may apply a board to ensure an increase in model transparency and rigor before models and solutions are fully developed and operationalized. This could involve more comprehensive testing and evaluation of the models and solutions to ensure that they are accurate, reliable, and fair, as well as more robust processes for addressing potential ethical concerns or unintended consequences. Additionally, it is possible that there will be greater public and regulatory scrutiny of the development and deployment of models and solutions, which could further drive the need for greater transparency and rigor in this space. Hopefully, part of this process will stimulate additional multidisciplinary colleagues to collaborate toward more ethical model building, data sciences processes, and overall decision–making practices.
7. COMPANIES WITH A STRONG MISSION-DRIVEN CULTURE WILL MORE SUCCESSFULLY ATTRACT TOP DATA SCIENCE TALENT
Mission-driven companies will have a competitive edge when recruiting analytics professionals. Through the immense volatility that the broader market has experienced over the last two years, the themes of work life balance and flexibility have come to the forefront with candidates seeking roles that allow them to work remotely or have added flexibility during their 9-5. To be competitive with today’s candidates, it is vital for organizations to have a long-term plan for flexibility and work-life balance. These have moved from “nice to have” perks to “must have” core requirements for many employees and candidates. Having a mission-driven purpose and company culture that supports these initiatives is vital for both attracting and retaining talent. With the difficulty in recruiting today, especially within the data science and analytics space, focusing candidates on a mission-driven opportunity rather than on a task-focused job is an effective strategy. Candidates are looking for Volunteer Time Off (VTO), a company that believes in diversity and inclusion, and one that focuses on the people. It is vital for companies to adequately convey their mission in a compelling manner to avoid difficulties with hiring when the market is competitive as candidates are increasingly prioritizing these factors when making career decisions.
8. DATA MANAGEMENT NEEDS WILL GET ACCOMPLISHED VIA AI METHODS
As user-centric data science becomes more prevalent, there is likely to be an increase in the application of data science toward data management processes such as feature engineering, data preparation, and metric generation. Automated systems like data observability can help to reduce the manual labor involved in data management by identifying and addressing issues in real-time. Additionally, the use of more models, including heuristics, to manage data needs is expected to become more common. With the increasing use of automated machine learning (AutoML) and other techniques for model development, the need for model-ready data will also increase in both quantity and quality.
9. COMMUNITY PLATFORMS FOR DATA SCIENCE WILL STIMULATE COLLABORATION ACROSS INDUSTRIES AND DOMAINS
Data science consortiums and other platforms will stimulate greater collaboration among leaders across multiple industries. The benefits of collaboration across industries such as CPG working with healthcare about regulatory requirements, or financial services data science leaders working with retail leaders about their consumer model. Aside from cross-vertical collaborations, we suspect the industry’s need for managing ethical AI and other advancements in ethically sustainable ways may serve as a forcing function to merge industries to meet to discuss common challenges.
10. DATA SCIENCE PRODUCT DEVELOPMENT WILL FOCUS ON USER CENTRIC DESIGNS WITH MORE HUMAN-MACHINE INTEGRATION
User-centric data science is an approach that prioritizes the needs and preferences of the end-user in the development of models and systems. This can lead to an increased use of automated systems and a shift in the skillset required for data science practitioners. Instead of a deep understanding of the mathematics behind the models, expertise in managing human-machine interactions and understanding the user’s perspective may become more valuable. This shift may also lead to a division of roles, with developers and software programmers from technology companies focusing more on the technical aspects of model development.