Original subtitles

Presenter: All right, so to kick off

the data modeling 101 part of the section,

you may very well be asking yourself,

well, what the heck is a data model anyway?

So I want you to picture something like this.

We've got three different tables shown here.

Product lookup, sales data, and returns data.

This is not a data model.

What we're looking at here is a collection

of three independent tables

which share no connections or relationships.

In other words, the sales data table

has no idea that the product lookup

or returns data tables even exist.

And what that means is that

if you tried to pull in or analyze data

from those sales or returns tables,

like the order quantity and return quantity for instance,

and filter those values based on some field

in the product lookup table,

like the product name,

this is what you'd see.

You'd get that same repeated grand total

in every single row of the table.

And again, that's because we have

no relationship, no connection

that's allowing us to filter

or segment those sales or returns values

based on fields outside of those tables.

Now, let's picture something like this.

Looks very similar, right?

We've got the same three tables here

but this is a data model.

And why is it a data model?

Well, because now we have relationships.

These three tables are connected

via relationships based on common fields or keys.

And we'll talk more about different types of keys

and what they mean in just a minute.

But for now, what this means

is that we can take fields

from our sales and returns data.

And because those relationships exist,

it means that we can aggregate

or filter those values

based on fields from related tables.

So now, instead of seeing that same repeated grand total,

we see accurate values

for these order quantity

and return quantity fields.

This is why data modeling is so so powerful

because it allows you to pull data

from all different places and sources.

And as long as you have valid keys

or connections to form relationships between them,

it means you can analyze that data

from all of the tables

in one centralized place

as if you're merging it all together

without having to physically join these tables

or write a bunch of complex formulas and functions.

So with that, let's go ahead

and open up our AdventureWorks report file in Power BI

and I'll show you exactly what this looks

like inside of the Power BI environment.

All right, so once you have Power BI pulled up,

go ahead and jump into the model view here.

This is where you should see all

of those tables that you loaded up

in Power Query in the last section of the course, right?

We've got our territories, our products, categories,

subcategories, customer lookup table, calendar table,

and our sales data table.

And remember, we haven't done anything

to create relationships between these tables

so we don't have a data model quite yet.

And to really hammer that point home

just like we did in the slides,

let's jump into the report view here.

We've got our blank canvas.

I'm gonna talk about

all these different visual types

later in the course, but for now,

we're just gonna use one type

of visual called the matrix visual.

You'll find it right here in the insert menu.

If you're using an older version of Power BI,

you may see it in a visualization pane

here on the right side of the canvas,

but it kind of looks like

a table with some blue shading on it.

So give it a click.

This will insert the matrix visual here.

And this works just like a pivot table in Excel.

We can pull in fields into rows,

into columns, and into values

to give us a nice little flexible table visual

that we can use to explore these numbers.

So let's go ahead and add data to our values.

We can either pull in

from the rightmost pane called the data pane

or we can do it right here in line.

Let's grab order quantity from our sales data table.

That's gonna automatically aggregate to a sum

and we see that same familiar total 84,174.

And remember, because we have no relationships yet,

that means that I can break down that total

using other fields inside

of the sales data table, right?

Like the product key for instance, that looks fine,

but what I can't do is break those values

down by any other table.

So to use the same example,

let's drill into our product lookup table,

bring in product name,

and this is where we see

those same repeating totals

row after row, after row.

So let's pause there.

That's kind of our 10,000 foot view

of what a data model actually is.

Next up, we're gonna talk about different types

of tables within a model

and how to create relationships

between them using primary and foreign keys.