Medallion Architecture Explained

So basically, medallion architecture, it's just a way to organize your data, usually in data lakes or lake houses.

It's architectural approaches in the big data world.

I will make a separate episode describing what data lakes and lake houses are.

But you'll just have to remember, overall, metaverse architecture is used in those approaches to organizing your data, so lake houses or data lakes.

It can be also used in a way in data warehouses, but then we have different terms and different layers which are most commonly used in data warehouses.

I will make a notion about it a bit later.

We have three layers in Medallion architecture.

So it's bronze, silver, and gold.

I don't know where they took the names from.

I guess maybe it comes from metals in the name of layers.

So I don't really know, but in any case, so it's bronze, silver, and gold.

And the difference between those layers, it's basically a way from the raw data, which is in a bronze layer, And then it goes and being processed, cleaned and prepared to silver and then to gold.

And in the gold layer, you can pretty much see it as a data marks layer.

That's the place where the business takes the data from to actually use this data.

So yeah, it's basically a level of organization, cleanliness, and all the stuff we need to do with the raw data to prepare it so the business can use it.

In data warehouses, we usually use the notion of a row or DS and data marts.

Those are the names of the layers.

So if you're familiar with these terms, you can imagine that medallion architecture is kind of similar, actually.

We can kind of connect that the raw layer is a bronze one, ODS is silver, and then gold is a datamars.

Although with silver it's a bit more complicated because it's not really like ODS.

But in general terms, you can kind of imagine that it's somewhat similar.

But it's also important to mention there is no requirement for us to have one silver layer.

We can have two.

We can make some processing and then on the next layer we do an additional type of processing.

And this is quite important for all the medallion architecture as far as I understand.