**Data modeling** is the process of creating a [[Data model|data model]] to represent the structure, relationships, and constraints of [[Data|data]] within a system or organization. It is a fundamental discipline within [[Data management|data management]] and [[Software engineering|software engineering]], used to define how data is collected, stored, and used in support of business processes and application development. Data modeling serves as a communication bridge between business stakeholders, [[Data architect|data architects]], and technical implementers, ensuring that systems accurately reflect real-world requirements. The data modeling process typically proceeds through three levels of abstraction: conceptual, logical, and physical. At the conceptual level, modelers identify key business entities and their high-level relationships without regard to technical implementation. The logical level refines this into detailed entity definitions, attributes, and [[Data normalization|normalized]] relationship structures. The physical level translates the logical model into technology-specific constructs suited to a particular [[Database management system|database management system]], including [[Table (database)|table]] definitions, [[Data type|data types]], and [[Index (database)|indexes]]. Throughout this progression, practitioners use formal notation such as [[Entity–relationship model|entity–relationship (ER) diagrams]], [[Unified Modeling Language|UML]] diagrams, and [[Dimensional modeling|dimensional schemas]] to document and communicate model design. Data modeling encompasses a range of techniques suited to different use cases. [[Relational model|Relational modeling]] is the dominant approach for [[Transactional database|transactional systems]], while [[Dimensional modeling]] is widely used in [[Data warehouse|data warehousing]] and [[Business intelligence|business intelligence]] contexts, typically employing [[Star schema|star]] or [[Snowflake schema|snowflake schemas]]. [[Data vault modeling]] offers an alternative approach for enterprise data warehouses that emphasizes [[Audit trail|auditability]] and adaptability to change. With the rise of [[Big data|big data]] and [[NoSQL]] systems, modeling practices have expanded to accommodate [[Document-oriented database|document]], [[Graph database|graph]], and [[Column-oriented database|columnar]] data structures. Emerging approaches in [[Data mesh|data mesh]] architectures treat data models as artifacts owned by [[Domain-driven design|domain]] teams, further distributing modeling responsibility across organizations.