Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Everything is obvious once you know the answer. The paperclip looks simple only because someone else did the hard work to invent it. The same applies to the Entity Relationship Diagram (ERD).
In the mid-1970s, databases were getting bigger, more complex, and harder to explain. Business users didn’t speak the same language as database designers. Developers struggled to describe data structures in a way that made sense to non-technical stakeholders. It is coincidentally the same problem we have 50+ years later.
Peter Pin-Shan Chen, then an assistant professor at MIT, saw the communication gap and decided to solve it.
In 1976, he published “The Entity-Relationship Model — Toward a Unified View of Data” in “ACM Transactions on Database Systems”
He wanted to create a universal visual language that could describe any data in a way humans could easily understand, before worrying about the technical implementation.
It was a really simple, elegant solution that consisted of three simple building blocks:
1. Entities – The things or concepts we care about (e.g., Customer, Order).
2. Attributes – The properties of those things (e.g., Customer Name, Order Date).
3. Relationships – How those things are connected (e.g., Customer places Order).
With this, a business analyst and a database engineer could point to the same diagram and agree on what was being modelled, long before a single line of SQL was written. Before Chen, data modelling often lived buried in technical schemas or scattered documentation. Chen’s ERD:
– Bridged business and technology: Everyone could see the same map.
– Abstracted away the storage layer: The model was logical, not tied to a specific DBMS.
It made data modelling visual, collaborative, and portable.
You might think a 1976 diagram style would be obsolete in a world of data lakes, object-relational mappings and cloud platforms.
But ERDs remain everywhere:
– Data vaults and star schemas still start with ERD-like conceptual diagrams
– Data governance platforms (like Collibra, Alation) use ERD-style views to connect business terms and physical data
– NoSQL and graph databases still benefit from entity-relationship thinking, even if the implementation looks different
– Cloud-native modelling tools (SQLDBM, ER/Studio, Lucidchart) still use Chen-style notation as a core visual language
The technology stack has changed, but the challenge Chen addressed, that being aligning people on what data means and how it connects hasn’t gone away. With data products, data mesh, and self-serve analytics becoming mainstream, clear, shared models are more important than ever.
You must be logged in to post a comment.