Secrets of Data Analytics Leaders

It's not easy being the head of data & analytics at a large organization. You must align a large team across multiple disciplines; you must deal with oodles of legacy systems and tools that hamper innovation; and you must deliver business value fast to keep executives at bay and your job intact. You also need to recruit dynamic managers who can push the envelope while meeting operational objectives. And when you falter--which you inevitably will-you have to rebound fast. No one knows these lessons better than Tiffany Perkins-Munn. She currently runs a 275-person data & analytics team at JP Morgan Chase that consists of data engineers, data scientists, behavioral economists, and business intelligence experts. She thrives on versatility, having earned a Ph.D. in Social-Personality Psychology with an interdisciplinary focus on Advanced Quantitative Methods. Building on this foundation, she has accumulated vast experience in the art of managing data & analytics teams during her 23 years in technical and managerial roles in the financial services industry. In this interview, you’ll learn: 1. Tiffany’s secret for aligning a large data & analytics team and keep them from splitting into silos of specialization 2. Her favorite techniques for recruiting the right people to her team. 3. How to wade through the thicket of legacy systems and deliver innovative solutions quickly. 4. The impact of GenAI on her operations and the financial services industry. 5. How to advance your careers in data & analytics.

Adopting community of practice principles, along with coaching and mentoring, is a practical approach to fostering and cultivating data literacy. Published at: https://www.eckerson.com/articles/a-people-first-approach-to-developing-data-literacy

This blog examines the upcoming trend of domain-specific LLMs and evaluates three different methods of implementation. Published at: https://www.eckerson.com/articles/the-next-wave-of-generative-ai-domain-specific-llms

Many machine learning (ML) use cases center on real-time calculations. This article defines streaming ML and its architectural components. Published at: https://www.eckerson.com/articles/machine-learning-and-streaming-data-pipelines-part-i-definitions-and-architecture

Companies need to invest heavily in teams and people, both at corporate and in the field, if they want to become a data-driven organization. Published at: https://www.eckerson.com/articles/organizing-for-success-part-iii-how-to-organize-and-staff-data-analytics-teams

Data management practices have changed substantially since the early 1990s and the dawn of data warehousing. Published at: https://www.eckerson.com/articles/the-continuing-evolution-of-data-management

Conventional data governance conflicts with today’s world of self-service analytics and agile projects. Published at: https://www.eckerson.com/articles/modern-data-governance-problems

Let's reflect on the events of the past year and prognosticate on what may transpire in the months ahead. Published at: https://www.eckerson.com/articles/trends-for-2024-our-team-gazes-into-the-crystal-ball

Data leaders must prepare their teams to deliver the timely, accurate, and trustworthy data that GenAI initiatives need to ensure they deliver results. They can do so by modernizing their environments, extending data governance programs, and fostering collaboration with data science teams. Published at: https://www.eckerson.com/articles/the-data-leader-s-guide-to-generative-ai-part-i-models-applications-and-pipelines

Data modeling is a core skill of data engineering, but it is missing or inadequate in many data engineering teams. These teams focus on moving data with little attention to shaping the data. They engineer processes, not products. Full data engineering is both process and product engineering, and that calls for data modeling. Published at: https://www.eckerson.com/articles/a-fresh-look-at-data-modeling-part-2-rediscovering-the-lost-art-of-data-modeling

The hardest part about implementing data products is fostering a product mindset among the people responsible for defining, governing, building, and shipping data products. It’s also important that an organization establish processes to facilitate the work of the product team and review boards. Published at: https://www.eckerson.com/articles/data-products-part-ii-data-products-require-product-thinking

Many organizations abandoned data modeling as they embraced big data and NoSQL. Now they find that data modeling continues to be important, perhaps more important today than ever before. With a fresh look you’ll see that today’s data modeling is different from past practices – much more than physical design for relational data. Published at: https://www.eckerson.com/articles/a-fresh-look-at-data-modeling-part-1-the-what-and-why-of-data-modeling

Data democratization is the buzzword to describe empowering enterprise stakeholders with data. While there have been advances in data management, governance, and analytics, something keeps getting in the way of achieving data democratization. Published at: https://www.eckerson.com/articles/data-democratization-and-the-duties-of-data-citizenship

Our industry’s breathless hype about generative AI tends to overlook the stubborn challenge of data governance. Data catalogs address this challenge by evaluating and controlling the accuracy, explainability, privacy, IP friendliness, and fairness of GenAI inputs. Published at: https://www.eckerson.com/articles/generative-ai-needs-vigilant-data-cataloging-and-governance

The need for an independent semantic layer continues to rise as data science gains traction in the enterprise. Its five primary elements—metrics, caching, metadata management, APIs, and access controls—support AI/ML use cases as part of data science projects. Published at: https://www.eckerson.com/articles/why-and-how-to-enable-data-science-with-an-independent-semantic-layer

Business leaders can address AI bias and use it to have rational discussions about management and human bias. Published at: https://www.eckerson.com/articles/weighing-the-risk-and-reward-of-ai-a-non-technical-guide-for-business-leaders

Most organizations view data as an asset to be actively managed with standards, controls, and discipline. Yet, they are passive and casual about metadata. Data is managed. Metadata happens. As data management becomes more complex, metadata management is becoming an essential discipline. It is time to think about metadata management from an architectural perspective. Published at: https://www.eckerson.com/articles/an-architectural-view-of-metadata-management

Kevin Petrie, the Vice President of Research at Eckerson Group, and Dan O’Brien, research analyst, discussed large language models (LLMs), which are neural networks that analyze text to predict the next word or phrase. These models use training data, often from the internet, to understand word relationships and provide accurate answers to natural language questions.

Generative AI initiatives require new data pipelines that prepare text files for querying by language models. Data engineers, scientists, and other stakeholders collaborate to design and implement these pipelines, which span text sources, tokens, vectors, vector databases, and LMs. Published at: https://www.eckerson.com/articles/the-new-data-pipeline-for-generative-ai-where-and-how-it-works

Dan and Wayne discussed the concept of data and analytics operating models, which refers to how organizations organize their data and analytics resources for alignment and efficiency.

This final blog in our series on the ROI of master data management recommends ways for data teams to iterate their MDM initiatives based on the successes and failures of their first project. Published: https://www.eckerson.com/articles/driving-roi-with-master-data-management-part-iii-project-iteration

Responsible AI ethical principles provide a clear, unifying purpose for the technological, business, and social goals of AI initiatives. Published at: https://www.eckerson.com/articles/the-opportunity-and-risk-of-generative-ai-part-iii-responsible-ai-ethics

Most definitions of a data product conflate it with a data asset. The only way to turn a data asset into a data product is to publish it in a data store along with metadata about subscription and delivery options, and terms of service that specify a bidirectional contract between data consumer and producer. Published at: https://www.eckerson.com/articles/let-s-be-clear-a-data-asset-is-not-a-data-product

Responsible AI can help data leaders comply with the fast-evolving regulatory environment of data and artificial intelligence. Published at: https://www.eckerson.com/articles/the-opportunity-and-risk-of-generative-ai-part-ii-how-responsible-ai-assists-compliance

US frontier history had races, risks, and rewards. Generative AI's future will follow a similar path. Published at: https://www.eckerson.com/articles/enterprise-data-and-the-taming-of-the-generative-ai-frontier

Dan and Jay discussed the concept of Data Fabric, an automated and AI-driven approach to managing modern data environments.

Simba Khadder and Kevin Petrie discuss strategies to overcome technical debt in implementation, the pivotal role of data in the success of ML projects, navigating regulatory compliance in machine learning, and the future of AI governance.

Learn how to attain an optimal return on investment (ROI) with MDM by choosing the appropriate architectural strategy and evaluating progress during the initial project implementation. Published at: https://www.eckerson.com/articles/driving-roi-with-master-data-management-part-ii-your-first-project

As organizations strive to meet the ever-growing demand for data, they are adopting data products to streamline delivery and ensure solutions provide value to business stakeholders. Learn about four traps that can disrupt data product development and how to avoid falling into them. Published at: https://www.eckerson.com/articles/four-traps-to-avoid-when-developing-data-products

Generative AI brings a promise to improve lives in a blistering innovation race, but also a threat to people, corporations, and even nations. Data analytics leaders must understand the risks of generative AI, both societal and business-related, to use it positively and avoid the destructive consequences seen with nuclear energy development. Published at: https://www.eckerson.com/articles/the-opportunity-and-risk-of-generative-ai-part-i-a-nuclear-explosion

The unbundling of the data ecosystem is causing organizations to “duct tape” products and frameworks together to build their solutions and data delivery processes. Organizations fail to build and deploy end-to-end, automated, repeatable data-driven systems, ignoring data engineering & dataops principles as well as best practices. Published at: https://www.eckerson.com/articles/dataops-in-data-engineering

This blog recommends guiding principles for successful implementation of language models to assist data engineering. Published at: https://www.eckerson.com/articles/should-ai-bots-build-your-data-pipelines-part-iv-guiding-principles-for-success-with-language-models-and-data-engineering

An emerging approach to generative AI will help data engineering teams achieve much-needed productivity gains while controlling risk. Published at: https://www.eckerson.com/articles/should-ai-bots-build-your-data-pipelines-part-iii-the-emergence-of-small-language-models-for-data-engineering

Dan O'Brien and Kevin Petrie discuss FinOps, which is a cost governance discipline for cloud-based analytics and operational projects.

An annual assessment of the positioning strategies of the leading 21 BI vendors finds a lack of differentiation that makes it difficult for buyers to compare products. In the BI market’s sea of sameness, Qlik is the only vendor that stands out with this clever, memorable position. Published at: https://www.eckerson.com/articles/independent-study-bi-vendor-messaging-shows-lack-of-differentiation

MDM creates business value in three ways: it streamlines infrastructure, streamlines processes, and reduces risk. Published at: https://www.eckerson.com/articles/driving-roi-with-master-data-management-part-1-build-your-business-case

“Universal” semantic layer tools introduced in recent years promise to standardize business metrics across the data stack, and eliminate silos of metrics trapped in semantic layers that are limited to specific data sources or BI platforms. This post offers considerations for adopting a universal semantic layer. Published at: https://www.eckerson.com/articles/the-universal-semantic-layer-more-than-enough

An Analytics Center of Excellence empowers business teams to meet their own data needs by changing the role of IT from developer to facilitator. The reality, however, is that IT needs be both a facilitator and a developer. Published at: https://www.eckerson.com/articles/analytics-center-of-excellence-part-i-how-to-shape-the-organization

Traditional companies must balance new and old technologies as part of an ever-modernizing data stack. This blog explores how companies strike the right balance to navigate economic uncertainty, AI disruption, and the need for tool consolidation. Published at: https://www.eckerson.com/articles/the-modernizing-data-stack-three-ways-to-balance-new-and-old

LLMs are hugely popular with data engineers because they boost productivity. But companies must adapt their data governance programs to control risks related to data quality, privacy, intellectual property, fai-Datarness, and explainability. Published at: https://www.eckerson.com/articles/should-ai-bots-build-your-data-pipelines-part-ii-risks-and-governance-approaches-for-data-engineers-to-use-large-language-models

Despite innovations in data architecture, infrastructure, and analytics, most organizations today still struggle to realize the promised value of data. Learn how the data mesh principle of data as a product can help, as part of a data mesh initiative or as a stand-alone strategy. Published at: https://www.eckerson.com/articles/data-products-part-of-a-data-mesh-initiative-or-a-stand-alone-strategy

Data mesh is a new paradigm for fulfilling the promised value of data. It decentralizes both data ownership and the data itself, shifting them toward the functional domains that create and use data to operate. But data mesh is not for everyone. Learn how to assess if you’re ready for data mesh. Published at: https://www.eckerson.com/articles/data-mesh-evaluating-your-organization-s-readiness-for-a-decentralized-data-future

There’s so much hype surrounding data products that you have to wonder if it’s just another buzzword. But there’s more to data products than buzz. In this article, you’ll learn how the concept is a meaningful step forward in the art and science of data management. Published at: https://www.eckerson.com/articles/best-practices-for-developing-and-scaling-data-products

A zone-based data refinery creates an agile, adaptable data environment that supports new and unanticipated business requirements quickly. It turns a monolithic data warehouse into a flexible data environment that gracefully adapts to new and unanticipated business requirements while maximizing reuse and standards. Published at: https://www.eckerson.com/articles/how-zone-based-data-processing-turns-your-monolithic-data-warehouse-into-a-flexible-modern-data-architecture

Many data engineers already use large language models to assist data ingestion, transformation, DataOps, and orchestration. This blog commences a series that explores the emergence of ChatGPT, Bard, and LLM tools from data pipeline vendors, and their implications for the discipline of data engineering. Published at: https://www.eckerson.com/articles/should-ai-bots-build-your-data-pipelines-examining-the-role-of-chatgpt-and-large-language-models-in-data-engineering

At IAPP Summit, privacy and data governance leaders expressed the importance of a collaborative operating model. Published at: https://www.eckerson.com/articles/the-convergence-of-data-governance-and-privacy-takeaways-from-the-global-privacy-summit

Embeddings are a learned way of representing data in space. Vector databases make it easier to work with embeddings generated from deep learning models. They will become an essential tool in the AI stack because they reduce the time to structure data and train models. Published at: https://www.eckerson.com/articles/the-why-what-who-and-where-of-vector-databases

A robust data workflow testing strategy helps ensure the accuracy and reliability of data processed within a pipeline. Use this checklist to meet your organization’s data quality requirements according to the dimensions of accuracy, completeness, conformity, consistency, integrity, precision, timeliness, and uniqueness. Published at: https://www.eckerson.com/articles/developing-a-robust-data-quality-strategy-for-your-data-pipeline-workflows

An operational data hub (ODH) is a pattern in data architecture that provides a central location and a standard protocol for operational systems to communicate about and share data among themselves. Operational systems post messages about data events (add, change, delete) and subscribe to messages of interest posted by other applications. The hub works to share data among applications without the clutter and chaos of point-to-point data feeds. Published at: https://www.eckerson.com/articles/operational-data-hub-responding-to-data-friction-and-technical-debt

Data mesh is a hot topic in the data world, generating conversations about the benefits and drawbacks of its decentralized approach. Concerns about an explosion of data silos and inconsistent data quality are justified. But to those who feel a bit like Chicken Little, maybe the sky is not falling. Published at: https://www.eckerson.com/articles/data-mesh-the-sky-is-not-falling