What is a semantic model in analytics?

A semantic model sits between raw tables and the people asking questions. It encodes what the data *means*: that a certain table is 'customers', that 'revenue' is a specific calculation, that customers relate to orders in a particular way. This is closely related to a semantic layer, the semantic model is the modelled content (entities, relationships, metrics); the semantic layer is the tier that serves it to tools and users.

A good semantic model provides:

• Friendly concepts business terms instead of cryptic column names.
• Defined metrics consistent formulas for key measures.
• Relationships how entities join, so queries are correct.

iDBQuery builds and uses a semantic model as a core capability. Its semantic layer learns the meaning of your tables and columns, even obscure ones, records the relationships between them, and lets you define business metrics once. That model is what allows you to ask in plain language, 'active customers by plan', and get correct SQL and a cited answer, without knowing how the schema is laid out. Because the model spans all your connected sources, the same friendly concepts can pull data from several systems at once, giving you consistent, understandable analytics over even a messy underlying estate.