Transcript

To keep their institutions competitive and resilient, data and analytics pros need to spot emerging trends and put them into action with smart strategies, tools and practices. In the 2025 Horizon Report Data and Analytics Edition, we look at six key technologies with the greatest potential to transform data and analytics in higher education. They include building institutional capacity for data literacy, AI powered assistance. AI powered decision intelligence. Mixed methods research for student learning data. Data mesh architecture and Federated data governance. Data literacy is the ability to use and communicate about data, understand how data are obtained and analyzed, and generate insights from data. Institutions that build capacity for data literacy are better equipped to generate insights aligned with institutional priorities. As one panelist put it, data is everywhere. It's information at our fingertips, particularly about the work we do and the people we serve. How can you take action? With great resources, all stakeholders can easily access develop guides, dashboards, glossaries and other resources to help users understand interact with data. Build a data literacy program that meets your communities unique needs. Make sure to on tighter your users level of comfort using data, their daily work and their modality and format preferences. Collaborate with professional learning staff. Staff who have expertise in designing effective professional learning experiences can help identify learning objectives, develop programming, and evaluate your data literacy program. Teach users to take a human centered approach to data literacy. Help your community use data for the benefit of all people, mitigating the risk of reinforcing biases and other barriers to success. Create a community of practice for data and analytics professionals or building institutional capacity for data literacy. Such networks help professionals share best practices and find opportunities to collaborate. Embed data literacy into the design, delivery, and governance of institutional data products. AI powered existence Our chat thought like tools designed to assist humans with digital tasks. They range from general purpose chat bots to specialized professional tools. All faculty and staff, including data and analytics professionals, can use AI powered assistance to help with daily work such as coding reports, preparing research briefs, and generating data visualizations. But there are risks. Outfits can cause hallucinations, i.e. erroneous or false details. They can also inadvertently expose sensitive information. The best AI powered assistants are not all equally available to all users, and differences in access can create or expand digital divides. One of our panelists said it best. One of the biggest risks we've seen is people putting too much trust in AI outputs, assuming it is always right or knows better than they do that. Kind of. Blind trusts can lead to bad decisions. So what actions can you take? Create internal documentation to help guide these effective and ethical use of AI powered assistance. Use data and AI powered assistance for institutional research In a great AI powered assistance into analytics dashboards, data users can ask contextual questions and receive on demand support. Identify workflow time syncs and match those pain points with available tools. Evaluate the efficacy of AI powered assistance. Have an evaluation plan ready even before procurement, and be ready to adjust implementation strategy as needed. Pilot and red team. Proper weaknesses New tools before releasing them to end users. Consider change management opportunities when integrating AI powered assistance. Focus on awareness, cultural acceptance, digital literacy training, and holistic support models. AI Powered Decision Intelligence comprises the host of tools, skills, and strategies that leverage AI for data, informed strategy, operations, and decision making. As AI powered decision intelligence becomes more mainstream, data and analytics professionals will have to decide how to position themselves in a new workforce landscape. Decision intelligence is shifting from reporting on what has happened to more contextual descriptions and explanations, including predicting future events and making suggestions for actions. There are risks with AI decision making. Key stakeholders such as students, faculty and staff could be left out of decision making processes. At best, poor data quality will render AI powered decision intelligence useless. At worst, it will lead to detrimental strategy and operations. Leaders can be tempted to put too much trust in insights generated by AI tools. As one panelist put it, AI is only as fair as the data and priorities that go into it, meaning there is a risk of quietly reinforcing existing biases, especially when no ones actively checking for them. So what can you do? Develop interdisciplinary working groups to guide AI powered decision intelligence. Continuously engage in professional development. Your institution will rely on you to know about cutting edge tools and best practices. Collaborate with business stakeholders to identify high impact decisions that can be enhanced with AI models. Invest time and budget for robust monitoring. AI tools must be routinely monitored to ensure transparency, trusts and efficacy. Put mechanisms in place to ensure data integrity. Low quality data, create low quality models and insights. Prioritize transparency, usability, and explain ability when selecting AI tools. Start with a solid foundation of data. Informed culture stakeholders should first feel comfortable with basic decision intelligence before engaging with AI tools. Mixed methods research for student learning data is a learning analytics approach that integrates both qualitative and quantitative research methods. Mixed methods research helps data users generate a more complete picture of stakeholder experiences by answering questions not only about how much or how many, but also about how and why. By supporting A wider range of research methods, data and analytics, professionals can foster stronger relationships and trust with more stakeholders. To quote a panelist, mixed methods research brings a more holistic understanding of student learning than quantitative or qualitative methods alone. It shows a human behind every number. Some of the risks can include stakeholders losing trust when qualitative data is an acted on skewed data and more sensitive and harder to anonymize data. So how can you take action? Collaborate with mixed methods researchers and other experts in mixed methods whenever possible. Includes stakeholders such as faculty, instructional designers, and students. Design studies that integrate learning analytics, survey data, focus groups, and more. Develop working groups or communities of practice with other researchers and data and analytics professionals interested in mixed methods research. Seek professional learning opportunities related to mixed methods research sample from a range of student experiences. Stay informed with the latest research on student learning. Start with reviews of past research and then regularly review new publications. Data Mesh Architecture is a distributed and decentralized model for data ownership and access. With a data mesh architecture, data are more likely to be owned by the right domain experts, leading to cleaner and more relevant data. Data and analytics professionals need to help their institutions adopt the appropriate mindset to use data mesh architecture effectively. More than just the Technology Strategy, it is a culture of cooperation and shared responsibility. However, without fostering A cultural shift that prioritizes collaboration, institutions that adopt data mesh architecture could create or reinforce fragmentation. Data users might duplicate efforts or even work against each other's interests if there is insufficient communication across units. As one panelist put it, while data mesh brings impressive promises for scale and innovation, it comes with its own set of challenges that are best tackled with open communication and a healthy dose of pragmatism. How can you take action? Develop institution wide standards for data through Federated data governance. Build continuous evaluation and improvement into data mesh architecture. This is not a system that can be set up once and trusted to run smoothly. And definitely partner with domain experts and when setting up data mesh architecture, embrace change. Data and analytics Professionals could be committed to data warehousing and other centralized models, but institutional change will require their buy in and collaboration. Seek professional learning related to data mesh. Provide training for data users. Stakeholders across campus need to understand their roles and responsibilities for owning, maintaining, and using data. Federated data governance is a hybrid model that enables institutions to provide some central oversight for data that are owned and controlled locally. Effective data governance is vital for ensuring that data are used appropriately and securely, and the centralized elements of Federated data governance ensure consistency for elements that are applicable to the entire institution. Federated data governance allows for more expansive participation and data processes, likely improving them. But Federated data governance is not a one-size-fits-all solution. Highly centralized institutions might not benefit from Federated data governance or might need to adopt it more gradually. What actions can you take? Start with solid data governance. Before attempting to transition to Federated data governance, be sure that your institution has a foundation of good data governance basics. Set up a shared governance framework. Create a collaborative governance council. Find the right tools. Implement enterprise level data governance tools that support automation, lineage tracking and policy enforcement. Be an advocate for data governance at your institution. Listen to your stakeholders, work together with representatives from all units, understand their needs, and collaborate on the creation of policies, guidelines and procedures. These insights are meant to help data and analytics professionals plan both short term moves and long term strategies to boost their institutions data maturity and impact. Explore the full 2025 Horizon Report Data and Analytics edition.