Original subtitles

All right.

So it's time for part two of our conversation about the power by key influencers visual.

Remember that part one we analyzed the impact to a categorical outcome.

So what influence is a Kickstarter project to be successful.

Now we're going to run a very similar analysis except we're going to use a continuous or numerical value

for outcome instead.

So in our power by a visual file I'm going to add a new page here.

I'm going to call it key influencers continuous now I'll just drop in a new instance of that key influencers

visual and we've got a nice blank slate to work with.

So just like before we're going to start by dragging a field into analyze.

But instead of using a categorical field like Project outcome this time I'm going to use a numerical

or continuous field like amount pledged.

So things look pretty similar here.

The only difference is that now we're saying what influences amount pledged to either increase or decrease.

Those are the standard options that we get with a continuous variable.

Now if you go into the Format tab and click into analysis you'll see that power b I detected this as

a continuous analysis type.

You can override that if you want.

I could say no no.

This is categorical.

And what we'll do is try to bucket every unique value in that field into its own category which is utter

nonsense.

So in this case you've got it right.

This is continuous and we're left with that increase decrease option.

So the question that we're posing here is what influences the amount pledged to increase to go up.

So let's start thinking about potentially influential factors here.

We know that the number of backers is highly correlated so we can pull that into the explained by well

here and check it out.

We get this nice scatter plot with dots representing each project in the table and we see this nice

clear linear relationship between the number of backers on the x axis and the amount pledged on the

Y which makes sense we have more backers more backers raise more money and we get that nice upward sloping

line.

This is a linear regression model and the slope of this line is what allows us to determine these factors

and these influencers.

So what we're saying is that when the number of backers goes up by nine hundred forty six the average

amount pledged increases by about one hundred nine thousand dollars.

Now here's where things get a little bit interesting which is that by default when you pull a continuous

variable into the analyze field you're going to see don't summarize.

And what that means is that power buy is going to run this analysis at the table level of granularity.

And what I mean by that is that the level of granularity in our source table is at the project level.

Each row each record represents one project.

That's why we see each dot in the scatter plot represent one single project.

But what if this is just too granular for me to even make sense of or generate insights that I can use

to optimize my business.

What if I actually want to see the relationship between backers and amount pledged at a higher level.

Like buy category or country or subcategory.

Well I can use the expand by option to do that.

What we can do is aggregate these values for both amount pledged and the number of backers that could

choose.

Average here instead.

Or it could take a measure.

In this case we've defined average amount pledged as the average pledged.

Makes sense.

We could use a measure here instead.

Same exact thing.

And we can choose an aggregation for our explained by fields as well.

So let's choose average for both.

And now what we're gonna do is determine the level of granularity that we want by pulling a field into

this expand by well.

So if we want to see the relationship between average backers an average amount pledged by category

we just grabbed category pull it in here and look at this.

We get that same scatter plot but now each dot no longer represents a project.

Each dot represents a category.

So the way to interpret this tooltip here is that on average projects in the games category have three

hundred and forty three backers and have raised about twenty five thousand five hundred dollars.

Same goes here with design projects with technology projects and so on.

So now on average for an average project when you get eighty nine more backers you can expect to raise

one point eight five thousand more dollars.

That's just a way to run a very similar analysis but explore things at a different level of granularity.

And when we pull a field into expand by what we're doing is determining that granularity without treating

it as an influencer which is what we'd be doing if we pulled that in here instead.

So there you have it.

That's how you can use power because key influencer visual to explore or predict a continuous outcome.