Skip to main content
SearchLoginLogin or Signup

A Framework for Phased Development of Academic Data Science

Published onOct 15, 2023
A Framework for Phased Development of Academic Data Science

A Framework for Phased Development of Academic Data Science

Table of Contents:

Introduction and Disclaimers

Phase 0 - Launching an Initiative

Phase 1 - Gaining Academic Recognition

Phase 2 - Build the Initiative

Phase 3 - Growth and Sustainability


This work was initially developed and presented by Jay Aikat (University of North Carolina, Chapel Hill) and Phil Bourne (University of Virginia) at the 2022 Academic Data Science Alliance Spring Meeting in Irvine, CA and further refined with Micaela Parker and Steve Van Tuyl from the Academic Data Science Alliance (ADSA) and input from the Data Science Leadership Community.


Over the past approximately two decades, we have seen the emergence of data science as a field and the subsequent proliferation of US and global academic data science degree programs and research programs established as part of new or existing schools, departments, institutes, or centers. We refer to these collectively as “initiatives'' to reflect their broad scope. As a relatively new and explosive field, and one that is inherently interdisciplinary, data science suffers from some of the same growing pains seen in previous emerging fields (e.g., computational biology) - precedents that may help guide university administrators and those responsible for establishing data science initiatives. As stated, some institutions have been building data science initiatives for over a decade, representing a range of possible strategies for creating and growing them. They represent a variety of US academic institutions, including universities and colleges designated as R1s, R2s, Minority Serving Institutions, community colleges and others - all supported by the Academic Data Science Alliance (ADSA), a non-profit dedicated to furthering the field.

It could be argued that data science is so fundamental to future scholars that it should be treated like mathematics, statistics, mastering a language, etc. Each of these fundamentals may be a distinct part of the institution’s organization even though it is practiced widely across disciplines. Hence, universities have mathematics departments, statistics departments and so on, but these subjects are also taught in other schools and departments. Arguably data science is no different and initiatives are evolving that are equivalent to what exists for these other fundamentals. It is an open and perhaps heady question as to whether data science is so pervasive as to subsume all other fields in some way given that digital data are available for all fields.

Based on real world experience, this document represents the ADSA community’s best understanding of the major phases of development for academic data science initiatives and key considerations for each phase. It is presented in the form of a checklist of opportunities and challenges. It draws upon five years of annual ADSA Leadership Summits where initiative progress has been actively discussed.

Briefly, we break the development of academic data science into four phases:

  • 0 - Planning and launching an initiative focused on education (e.g., a specialty, certificate or degree), research, and/or consulting;

  • 1 - Establishing an academically recognized data science unit (e.g., institute, center, department, school) likely including interdisciplinary research;

  • 2 - Building out that unit, gaining academic recognition; and

  • 3 - Achieving manageable program growth and sustainability.

While these phases are a helpful framework for considering the trajectories of data science initiatives, we recognize that this framework may not exactly align with the needs of any individual institution. However, we believe that providing these major milestones for the development of academic data science can provide a common language for understanding the issues and challenges academic institutions will face along the trajectory of initiative development and growth.

For each phase, we consider the vision, mission and goals, the financial model, organizational structure, and governance models of the initiative. Importantly, we also consider how diversity, equity, and inclusion can be built into each phase.


There is so much variability in how higher education operates. Even within similar institution types (e.g. 2-year, 4-year, public, private, etc.) variability in organizational structure and administrative culture make each institution unique. This framework undoubtedly falls short of covering the entirety of this landscape, but will be a valuable element of the ensemble of resources at hand.

Likewise, every institution has its politics and personalities, some of which will likely be engaged in the process of launching and growing data science initiatives. These relationships, collaborations, and “backroom deals” will be a necessary part of nurturing a data science initiative.

Some points of note carry over from one phase to another and are listed multiple times. That redundancy is intentional.

Phase 0 - Launching an Initiative

The first step to formalizing a data science effort on your campus is typically the creation of an initiative that serves the campus community in some way, often in response to workforce needs, faculty-driven educational goals, research coordination needs for grand challenges, or university leadership’s institutional goals. At this phase, you are likely having discussions across multiple units in the institution, a result of the interdisciplinary nature of data science. As such this will likely define the future of data science at the institution. Your data science initiative could be centralized or distributed across multiple schools and departments, affiliated with one or more schools and departments or a new stand-alone academic enterprise. It is too early to predict if one model works better than others. It can be said that if the data science initiative is affiliated with an existing school or department its flavor and culture will be that of the existing unit. Whatever the model, a sense of confusion and territoriality will likely occur and will require strong academic leadership to chart the best path. Without such leadership there is a danger that your organization will not be prepared to build a coherent data science initiative, let alone be ready for the rapid changes catalyzed by large amounts of digital data. Decisions by institutional leaders may be shaped by the institution's specialities, philanthropic donations, and politics.

Deciding what to focus on first: education or research support

What is your vision for data science at your institution and what initial organizational structures need to be in place to build toward that vision?

  • Start by doing a bit of institutional and inter-institutional research to understand the demand for data science educational programs versus research consulting and support programs across campus. The demand will undoubtedly be there, the bigger question may be what to focus on, and thus what not to attempt, at least initially.

  • A major decision point is whether to focus initially only on education (academic degrees and certificates) or stand-up a data science institute or center to promote and support data science training (non-academic, e.g. for staff) and research consulting for the campus community.  

  • Internally data science is likely considered both an opportunity and a threat to existing disciplines and scholarship. Consider the political / territorial landscape of current efforts and how your new initiative will add benefit and impact to current programs. Meet with those other programs to build relationships and minimize in-fighting while jointly taking advantage of demand.

  • What impact will a new data science initiative have on existing programs and departments? Understanding how a new data science initiative interfaces with existing teaching and research will likely be paramount to the success of the new data science initiative and also to the sustained success of these other existing programs and departments.

  • Beyond the demand, are the human resources and the physical infrastructure in place to launch an initiative? Human resources includes not only faculty but support staff, both within the new initiative and also across campus in, for example, student support.

  • For instruction, consider online versus residential, and indeed since COVID, support for hybrid instruction.

Determining what is unique about your program

Data science programs are common at this point, what gives you a competitive edge against peer institutions? What will your mission and goals be?

  • How to differentiate your institution’s data science initiative:

    • Start with an institutionally shared definition of what data science means to your institution. This is likely achieved through significant discussion with faculty across the institution. The shared definition will likely morph over time, but it gets all stakeholders engaged and on the same page.

    • The definition of data science may include a focus on one or more domains/verticals depending on what your institution is known for.

    • Data science offers an opportunity to strengthen ties to those in industry, government and NGOs. Consider what is possible within your geographic region and beyond.

University commitment

This commitment depends on the structure of the university but typically comes from the university leadership, the faculty, through the faculty senate/council, and the state for public institutions. If the initiative will include new buildings etc. it may involve local authorities.

  • Your institution likely already has a strategic plan designed to differentiate itself. Consider how your data science initiative can align with that plan.

  • Understand how your institution makes strategic decisions - top-down, bottom up.

  • Identify champions who have the ear of leadership. Use these individuals to start a form of governance for the initiative.

  • As much as possible, work with the leadership, president/chancellor, provost, deans, etc. Their support will be critical. They will likely ask questions like:

    • What are other institutions doing and are those efforts successful?

    • How much will it cost?

    • What might my school be asked to give up in the process (in the case of deans worried about sharing strained institutional resources)?

    • Is it sustainable?

    • Do other stakeholders want it? Recognizing that different stakeholders will be driven by different motivations. For example deans by financial considerations, faculty by a consideration of their reputation.

    • What career paths does this degree or initiative open that we do not already support?

  • Be prepared to address the above questions. Your answers will also evolve as the planning moves forward. The economic question is particularly important. Extramural funding and/or philanthropy is an important lever if available. Startup funding is essential; developing and working toward a sustainable business model is key.

  • Determine the ability to bring in extramural funding and philanthropy to bolster budgets as needed.

  • Consider how other new initiatives on your campus have fared recently and learn from them; how they were influenced by the culture and governance of your institution.

  • Consider starting small: Identifying one modest initiative that could be launched successfully and used for developing broader support and resources. Initial successes help in cementing longer-term plans.

Challenges in Phase 0

  • University change is much slower than the evolution of data science.

  • Determining the order and timing of the launch of different elements of the initiative (e.g. degree programs, consulting services, etc.) - some institutions may be more amenable to incremental changes, while others may be open to more sweeping changes.

  • How to establish a brand beyond the initial announcement.

  • Finding ways to compromise and make the effort succeed despite setbacks. It is good to have ideas of backup plans and/or separate program elements that can succeed independently of others.

  • Not to own everything but be seen as a partner in a bigger enterprise - a collaborator rather than a threat. This includes acknowledging that various other units may already have courses, curricular programs, and/or research programs in data science. Therefore:

    • Collaborate with various academic units at the same university starting data science programs.

    • Work to admit more students rather than compete for the existing pool.

    • Avoid redundancy and inefficient use of institutional resources notably space, compute and human resources. Working through this will help you develop an initial financial model.

    • Share what resources you do have.

Opportunities in Phase 0

  • Beginning to fundamentally change your institution in positive ways that serves all stakeholders.

  • Recognizing and leveraging the existing strengths at your institution in data science across not only the STEM disciplines but also social sciences and humanities. Arguably this recognition will be embraced for financial as well as the wellbeing of society.

  • Creating a sense of community and camaraderie across your institution around data science and building a new initiative with values and commitments to diversity, equity, inclusion, justice, and accessibility.

Phase 1 - Establishing the Initiative

You have started a data science initiative, what comes next? Data Science initiatives likely coalesce towards an academically organized data science unit, such as a department, college, school or institute, among others. Forces that propel this growth may include increased demand for teaching or an increased demand for research or research support services or simply a desire to put data science on a par with other forms of scholarship based on market forces. Importantly, now is the time to revisit your vision statement and your mission/goals to ensure you are not getting pulled too far from what you had determined was the best fit for your institution, or revise these to reflect the changing environment and needs of your campus.

Phase 1 likely starts around year 3 of the initiative.

Establishing the unit with minimal organization, processes, and resources

As your initiative evolves into a broadly recognized data science unit on campus, the organizational structure of the unit will naturally evolve. What will this look like and who will be involved in the governance?

  • Is it better to start from scratch or merge with existing units/initiatives?

  • Pros and cons of each approach

    • From Scratch

      • Pros: Opportunity to shape something that is unique in higher education. Address DEI from the outset; opportunity to not follow “this is how it’s always been done” route and challenge “but it’s never been done like that here” comments.

      • Cons: Takes enormous vision, time, institutional support. Lots of upfront staffing for student services, financial services etc. Finding the faculty to teach and incentivizing them. Most universities have not launched new units of this scale in decades and thus there is no playbook to get this done methodically. This often means that even where there’s strong institutional support for the initiative, asking the right questions and seeking help from the right people on campus becomes key.

    • Merge Existing

      • Pros: relatively light lift; have a ready pool of faculty and staff to begin; institutional / organizational infrastructure is in place; leveraging existing courses more easily for new degree programs in data science; seen as less of a threat since it’s not a new unit at the university’s limited resources table.

      • Cons: competing interests; merging of cultures to form the new unit can take years and much work; allocation of new data science resources may get diverted to old starving problems.

  • Whether to start a unit with research, education, or both 

In Phase 0, you had to decide whether the new initiative will focus first on education or research. Now, in Phase 1 you have taken the pulse of the community, established demand for your choices (or lack thereof) and can decide to lean into formal establishment of the initiative as an academic unit (department, school of data science) or a cross-university research and support institute, or recognize that a previous assumption was incorrect and it’s time to change course. In some ways, this is the point of no return. You are all in. This doesn’t mean you can’t add-on other programs to your initiative later, but your investment now is in formalizing the choice.

  • As described above, the choice depends on the overall university emphasis, on the revenue needed, which in turn is related to the university funding model, and  on the current programs or courses available across various existing units, delving into data science within their current domains. 

  • Education Focus

    • Where to start? professional (graduate), undergraduate, or continuing education.

    • Does your approach meet the needs of employers, and the state (especially for state institutions)?

  • Research Focus

    • Does the research map to what the institution is known for?

    • Are there sufficient research faculty and/or research staff to have an impact?

  • Decide if both education and research will be co-located, if not, how will you coordinate them.

  • Develop and have a strategic plan in place (vision, mission, goals) that is dynamic, has defined metrics, and aligns with the University’s strategic plan and was developed by a comprehensive group of stakeholders.

    • Determine what guiding principles apply to the plan, e.g. quality, openness, transparency, translation, inclusion, diversity, equity, access, and weave those principles into all aspects of the plan.

    • Strategic planning helps solidify the areas of focus for the unit and gets all stakeholders on the same page.

    • Whether starting a unit from scratch or a merger, seek input and feedback from campus stakeholders during, and not just at the end of, the development of the strategic plan.

    • Have the plan dovetail with any university-wide strategic plan.

    • Have the plan address the culture you want the unit to have.

    • Formation will likely be part of a larger process of formally implementing the unit at the institution level

    • Have DEI as an integral part of the plan - not as an adjunct, but woven into all the unit does; commit real resources to implement the DEI parts of the plan.

  • Provide clarity for how the unit is unique both within and outside of the university.

  • Seek input through advisory councils from within and outside the university; the kinds of such groups and their memberships may depend on the culture and governance at your institution.

  • Continue the phase 0 differentiation process. Formalizing an initiative will require continued differentiation to help administrators (in the program and elsewhere at the institution) understand why the program is needed.

Challenges in Phase 1

  • Deciding what not to do.

  • Thus, not having too few people doing too much leading to burnout.

  • Obtaining buy-in from other units – making the case – what’s in it for them and how much is it going to cost.

  • Establishing a financial model of operation that promises to be sustainable, including one-time and recurring costs.

  • Standing up clear processes.

  • Engaging various academic units at the same university starting DS programs

  • Hiring - faculty, researchers, data scientists, software engineers, and administrative staff in a competitive market.

  • Dealing with the complexity of joint and affiliate appointments; different but equally challenging for new assistant professor joint hires as for making existing full professors joint between two units.

  • Evolving responsibilities as found in a startup-like environment - employees have to be comfortable with rapid change.

  • Professional development for all employees in a fledgling and fast changing organization.

  • Hiring a diverse faculty and staff who are committed to diversity, equity, inclusion, and access.

  • Creating an inclusive and welcoming environment.

  • Providing sufficient physical and virtual resources - space, infrastructure, people.

  • Suitable provision of student support and wellbeing services.

Opportunities in Phase 1

  • Start to build your own brand

  • Increase visibility and goodwill across campus.

  • Fundraising opportunities.

  • Meeting workforce goals.

  • Conducting interdisciplinary research not typically achieved in a siloed organization.

  • Develop and maintain a public “dashboard” that aligns with key objectives of the strategic plan.

Phase 2 - Build the Initiative

Once your data science initiative has a foothold the next phase is to continue to provide resources (personnel, funding sources, program elements) aiming at sustainability. This phase will test the strategic plan and other guiding principles. Course corrections will be necessary. Leadership will need to make a number of hard choices about the focus of the initiative (teaching, research, consulting, etc.), how the initiative engages with other units at the institution, and how the initiative will engage with communities beyond the institution, such as the local or regional community, government, or industry. At this stage, governance may also change. It may be a good time to rotate out some of your executive/steering committee members for fresh energy, and start an external advisory board, if one hasn’t been established already.

Phase 2 likely begins in years 4 or 5.

Implement the Strategic Plan from Phase 1

  • Identify intermediate goals for the initiative based on the plan.

  • Evaluate - Identify measures of success and report on them to all stakeholders - university leaders, funders, faculty, staff and students through a dashboard kept current or other means. Do this early, better in phase 1 so there’s a baseline against which to measure progress.

  • Strategic plans are meant to be working documents - don’t fear flexing the plan to ensure the initiative’s needs are being met.

Institutional Alignment

  • Understand the role within the larger institution (what will data science mean in 10 years and how does it fit into the institution, does it exist as a separate entity or integrated?).

  • Does the initiative award tenure? If so, how does it align with tenure policies within the rest of the institution? Similarly for how promotion is handled.

  • Recognizing that the institutional strategic plan may evolve, re-align, as needed, the initiative’s strategy with the institution-level strategic plan.

  • Get a seat at the leadership table where long term planning discussions are happening. This can take various forms depending on the organizational structure of your institution. Attending deans meetings, faculty senate representation, university-wide strategic planning committee membership are possible examples.

  • Work to leverage university growth to align with your own needs for physical space, compute resources, human services etc.

  • Find avenues to lead and provide institutional resources and thus win wider support in the university; e.g., shared datasets for classes, shared resources (people and infrastructure) to facilitate development of large proposals and implementation of large projects across disciplines.

  • Consider how your interdisciplinary data science initiative naturally interacts with other interdisciplinary programs on campus. Are there lessons that can be learned from these other initiatives? Are there spaces to collaborate?

Expand Offerings

  • Launch additional degree programs, consulting efforts, research consortia, industry partnerships ensuring the human and other resources exist to carry them out.

  • Determine how/if these align with strategic plan and program resources.

  • Start an external advisory board to help make decisions on what will be useful to the university community (and beyond) and support the growth of the data science unit; advisory boards could be external to the data science initiative - i.e., campus stakeholders representing other units, or external to the university - i.e., industry or community partners.

Research Specialization

  • Consider specific research directions in data science across the institution.

  • Consider the economics of such research which will depend on the institutional funding model.

  • Identify whether a specific research domain appears to be a focus for the program and/or if there are existing research groups, labs, institutes on campus that hold synergy with the program.

  • Use such research directions and existing faculty in these areas to (a) form large clusters (could be formal or informal) that could be supported to pursue large projects; and/or (b) hire new faculty and research staff to augment in these areas.

Community Engagement

  • First, identify the community - is this the local (geographically local) community? A specific research area or industry community? Other?

  • Determine the level of community engagement, including industry

  • Establish workforce pipelines as appropriate.

  • For all of these, lean into your values of diversity, equity, and inclusion. Be deliberate in raising the voices and opportunities for those who are under-served in your communities.

Seek funding

  • Consider whether hiring a development person is warranted; or advocating for representation with the institution-level development office.

  • Consider the role of gifting, corporate sponsorship/partnerships and the emerging alumni.

  • Consider the role of foundations and government agencies in funding.

  • Determine the role of unit development with respect to development institution wide.

  • Consider the role of the advisory board in funding.

  • Consider the relationship with your government relations group as it relates to local, state and federal governments.

Challenges for Phase 2

  • Managing rapid growth

    • Again, deciding what not to do

    • Hiring fast enough to meet objectives

    • Obtaining the necessary physical resources

    • Getting processes in place fast enough, without sacrificing clarity, robustness, and shared governance (e.g. with faculty and administrators)

    • Maintaining morale in a high pressure environment.

  • How to continue to advocate for and fund:

    • FTEs both faculty and staff.

    • Physical space.

    • Compute resources.

    • Marketing, communications, IT services, development, DEI, etc.

  • Achieving a balanced budget.

  • Retaining faculty and staff in a competitive environment.

  • Providing professional and career development for faculty and staff.

  • Implementing the appropriate student services.

  • Preparing for alumni.

  • Further building of the brand (aka creating an Identity).

  • Education and research programs will evolve fast. How to keep your unit's position within the institution and beyond.

  • Beyond simply differentiating the program from others (inter-institutional and intra-institutional differentiation); address (a) the anxieties faced by some other units that are threatened by your existence and growth as also (b) the excitement expressed by some other units at the potential for successful new collaborations. And recognize this will be an ongoing process.

  • Identify strengths of the organization and its people.

  • Identify emerging growth areas and determine if resourcing allows focus in these areas.

  • Is the initiative’s focus methods, applications, a combination, addressing pressing societal problems, if so which ones - what is emerging? Is it consistent with the initial vision/mission? What course corrections are needed?

  • Identify research foci for the next 5 years with a deliberate process of building upon existing strengths at the university vs adding areas that are already lacking.

  • Hiring the right people at all levels and all areas of the organization (critical).

  • Building deliberate processes and systems (which will evolve) for professional development and mentoring of faculty and staff, paying attention to DEI and access to opportunities for all.

  • Creating a balanced revenue portfolio - tuition, F&A, Public-Private Partnerships (PPP) (e.g., workforce partners), philanthropy.

  • Assessing whether your definition of data science remains appropriate or needs revision.

  • Having the right constituency in your advisory board.

  • Exercising appropriate caution when engaging with industry partners?

  • Grappling with salary discrepancies within the unit based on the market for salaries in the different disciplines that data scientists come from; this is especially true for faculty hires, but also in some research staff roles. The career guide Hiring, Managing, and Retaining Data Scientists and Research Software Engineers in Academia: A Career Guidebook from ADSA and US-RSE might help!

Opportunities in Phase 2

  • Recognizing what you want to be known for in 10 years and starting to go for it.

  • Begin to consider larger partnerships through federal funding, state workforce development programs, public private partnerships etc.

  • Consider how to leverage your unit to serve as a catalyst / hub for large multidisciplinary data science projects.

  • As courses are developed, there are multiple opportunities to tag-up with existing degree programs to enable X+DS degrees in various fields. Likewise for certificates, minors and lifelong learning activities.

Phase 3 - Growth and Sustainability

The last phase approaches a steady state - educational programs are in place, research areas have been identified with grants secured, and industry and community engagement is underway. The goal is to ensure a sustainable funding scheme, and address changes in the field of data science and how they may impact the unit. Continual revisiting of the vision and mission statements, and the governance will also be crucial at this stage.

By way of example, the advent of large language models (LLMs) is an example of rapid change. What is the role of the initiative in the fast changing world of AI? How much belongs in the initiative, how much resides in other parts of the institution. How to reconcile these different developments to avoid confusion among stakeholders and do the best by the institution.

Progress is likely not linear as the field of data science grows. Some aspects will be found by returning to earlier phases and rebuilding anew. In short, an agile approach is needed when developing the initiative.

Phase 3 likely begins in year 7 onwards.

Metrics of Success

  • What to measure to determine if your program is meeting stakeholder expectations or not. These may not only be traditional metrics, e.g., papers published, research grants received, students trained, but other aspects in alignment with the strategic plan and guiding principles. Examples might be evidence of local community engagement, software and data used by others, etc.

  • Evidence that your unit has shifted the needle in how higher education is thought of to be more in alignment with workforce and research needs.

  • Accreditation, depending on your institutional accreditation model, as well as that of the emerging field of data science. If accreditation specifically for data science degrees is desired, there are currently two programs within ABET that offer this for data science, each with its own flavor: the Computer Science Accreditation Board (CSAB) and the Applied and Natural Science Accreditation Commission (ANSAC). Accreditation for data science is in early days, and we encourage data science programs to keep an eye on this landscape. It is anticipated that accreditation will conform to a snowball effect. Once a few institutions go down the accreditation path many others will follow in an effort to compete. The true value of accreditation is hard to judge and depends somewhat on the type of stakeholder. For example, does industry see it as important that the students they hire are from an accredited program?

Organizational Positioning

Where does your unit stand according to your own stakeholders and those not affiliated.

  • Internal

    • A clear picture of where data science fits in the overall organization.

    • Alignment of strategic plans - institutional and unit.

    • Integration with other units that leads to a win-win situation.

    • Ensure the unit remains a priority for the overall institution.

  • External

    • Feedback from an external advisory board.

    • Accreditation.

    • Identifiable differentiators from initiatives at peer institutions.

Sustainable Funding Model

  • Evidence of acquiring core, recurring funding.

  • Funding that is ethical and supports the unit's mission.

  • Survival through prior period(s) of financial instability.

  • Provision of student fellowships.

  • More specifically, ability to provide financial assistance, particularly to first gen students and those socioeconomically disadvantaged.

Organizational Agility

  • Unit can weather personnel loss and other perturbing influences.

  • Unit changes as the definition/perception of data science changes.

  • Thoughtful reflection on changes in the field - avoiding knee-jerk changes.

  • Adjusting policies and procedures to match the changing field, notably related to promotion and tenure. Do the policies map to the expectations of the field and yet are not too distant as to not be recognized by other disciplines. There is a fine line between pushing forward a new model of what scholarship is valued and what is seen as improper by other fields.

  • Unit continues to engage a diverse set of stakeholders - staff, students, faculty.

  • Other units building programs, hiring faculty, or building infrastructure in data science are seen as synergistic rather than competitive because of a solid foundation of finances and resources and explicit connections to these other units (shared seminar series, community events, jointly funded fellowships, for example).

  • Development of the workforce pipeline outside of higher education, i.e., with K12, lifelong learners, and community college to 4yr degree pathways.

Challenges in Phase 3

  • Maintaining the momentum of earlier startup phases.

  • Retaining the early culture and values in a larger organization.

  • Maintaining appropriate communication in a larger organization.

  • Not becoming obsolete as other units adopt data science.

  • Changing hiring practices to be more equitable, growing diversity and ensuring a welcoming environment for all.

  • Continuing to be creative and agile as the unit grows.

  • Making joint hires work in a way that is equitable to the other collaborating units.

  • Balancing the budget, possibly without ongoing institutional support.

  • Maintaining a philanthropic base when not the new shiny object.

  • Nurturing and utilizing the growing number of alumni.

  • Applying data science to yourselves to achieve continuous improvement.

  • Still deciding what not to do in a continued myriad of opportunities.

Opportunities in Phase 3

  • Achieving recognizable impact that influences our emergent digital society in positive ways.

  • Impacting policies at regional, state, and federal levels.

  • Training generations of workers that map to the needs of a high-demand society.


Data science is being adopted broadly across the academy. Developments vary from specific courses, complete educational programs awarding degrees, research institutes, consulting groups, departments to whole schools. These developments are in response to the digital transformation of society requiring the academy to train the next generation workforce, upskill the existing workforce, and conduct research that uses the data and tools of this transformation to better our world at local, regional, national and global scales.

Data science adoption has existed in some form in the academy for at least two decades even as the rate of adoption has increased. What have we learnt along the way that is relevant for those responsible for starting or further developing data science initiatives? We have tried to answer this question by considering adoption in phases 0-3 which seems to reflect what is happening across the academy. Each phase presents a checklist of opportunities and challenges some of us have faced along the way.

As a field where supply is far outweighed by demand and being relatively new without an entrenched culture, it is refreshingly open and transparent. The thoughts presented here are intended to support and further that notion.

No comments here
Why not start the discussion?