Techsplainers by IBM

This episode of Techsplainers explains data reliability—the completeness and accuracy of data across time and sources. We explore how reliability is measured through validity, completeness, and uniqueness, and distinguish it from related concepts like data quality and validity. The discussion covers common challenges organizations face with data reliability, from collection methods and human error to changing sources and duplication issues. We provide practical steps for ensuring reliable data, including standardized collection, proper training, regular audits, and strong governance. Finally, we examine how data observability transforms reliability management by enabling real-time issue identification and resolution before bad data impacts decision-making. For organizations seeking competitive advantage through data-driven decisions, establishing robust reliability practices is no longer optional but essential.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Mimi Sun Longo

This episode of Techsplainers explores data quality dimensions – which provide the structured framework for measuring and evaluating data trustworthiness. We explain the six core dimensions: accuracy (correctness of data), completeness (presence of all required values), consistency (uniformity across systems), timeliness (currency of information), validity (conformity to rules), and uniqueness (absence of duplicates). The episode delves into why these dimensions matter – with poor data quality costing organizations millions annually – and outlines a three-step implementation process: assessment, measurement, and continuous improvement. We also highlight key benefits, including better decision-making, regulatory compliance, workflow optimization, customer satisfaction, and risk reduction. These dimensions provide the foundation for trusted data that powers reliable insights and automation.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Mimi Sun Longo

This episode of Techsplainers explores data quality—the measure of how well datasets meet criteria for accuracy, completeness, validity, consistency, uniqueness, timeliness, and fitness for purpose. We examine the seven key dimensions of data quality and explain how they impact business decisions, processes, and customer satisfaction. The discussion highlights the critical distinction between data quality, data integrity, and data profiling, while explaining why poor quality data costs organizations an average of USD 12.9 million annually (according to Gartner research). We also explore the growing importance of data quality for AI and machine learning systems, where the "garbage in, garbage out" principle directly affects outcomes. Whether you're in marketing, supply chain management, or healthcare, understanding data quality fundamentals is essential for making reliable, data-driven decisions.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Mimi Sun Longo

This episode of Techsplainers explores Code LLMs, specialized AI models that are transforming software development by understanding, generating, and explaining code. We examine how these models differ from general-purpose LLMs through their extensive training on programming languages, documentation, and code repositories, giving them a deeper understanding of software development concepts and patterns. The discussion covers how Code LLMs work through transformer architecture and reinforcement learning from human feedback, as well as their substantial benefits for developers, including increased productivity, enhanced learning opportunities, improved documentation, and democratized access to coding. We highlight practical applications such as code generation, bug fixing, refactoring, and test creation, while also addressing important limitations including potential security vulnerabilities, challenges with complex code, intellectual property concerns, and the risk of over-reliance. The episode concludes with insights into how these tools are reshaping the development landscape, with developers increasingly shifting toward a supervisory role focused on architecture and design decisions.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Erika Russi

This episode of Techsplainers explores COBOL modernization strategies for organizations managing critical legacy systems written in this 60-year-old programming language. We examine why COBOL remains vital today—processing 95% of ATM transactions and 90% of global financial transactions—while highlighting the pressing challenges of maintaining these systems amid a growing skills gap, integration difficulties, and scalability concerns. The discussion covers various modernization approaches, from less disruptive strategies like rehosting and refactoring to more transformative methods like rearchitecting and replacement. We share key best practices for successful modernization, including thorough application documentation, business-value prioritization, preservation of essential business logic, and incremental implementation. The episode emphasizes that effective COBOL modernization isn't about wholesale replacement but thoughtful evolution that preserves decades of refined business logic while enabling future innovation.

Find more information at https://www.ibm.biz/techsplainers-podcast

Narrated by Erika Russi