Everybody.

And welcome back to another data science with python and pandas tutorial video In this video, we're gonna be building on everything up to this point.

We're gonna continue with the minimum wage data set for one more video.

This one, uh, and kind of combined to other data sets just to Seymour about this whole relationships thing and learn a fume or cool features of pandas before I finalize this little miniseries with doing machine learning.

Ah, basically a machine learning workflow with pandas in the next video.

With that, let's get into it.

So I'm gonna be using to new data sets here.

We're gonna be using the 2016 presidential election vote by county as well as unemployment by county.

So pretty cool.

So it be nice to combine all of these data sets into one major data set and see if voting Proclivities, the unemployment rate and the minimum wage have any relationships with with each other, Right, Cool.

So that's the idea on really we're just going to see if there's any relationship.

It's pretty hard to based on that data loan, draw any riel conclusions, we're just doing it to c t learn more about pandas were really So if you find yourself offended by the results of this video um, not my problem.

So anyways, here we go import, import, import pandas as p d.

And then we're gonna bring in unemployment county as p dot Read under scorer Sissi And that is data sets slash.

And this one is unemployment by county dash us slash output that CSP.

So this one came with his own little directory, basically.

And then, um, we could do on an county dot head.

Okay.

And there we get a rate that's our unemployment rate on.

Then it's by county, and then you're so cool.

So now we want to do is combine this plausibly with our minimum wage.

So to do that, we're just gonna come over here and basically take all of this data Copy.

Come over here, paste Boom!

Awesome.

Okay, so now we want to do is ah, we saw that missing data.

So let's go ahead and fix that real quick.

Um, in fact, can I just do this?

Yes, I shall.

Working.

No, wait.

I just want to do this.

I don't want the core.

I just want to drop it.

So instead we'll say Atzmon wage equals act min wage, uh, dropping in and then act me had fixed that capital w Okay, great.

So now what do we want to d'oh.

So we want to take the minimum wage and we want to kind of add that to a new column in our unemployment county data set.

So the way I think I want to do that is by creating a new column and then mapping the values.

There's probably a better way than this, but it gives me the opportunity to show you guys at the most basic level a method that will always work for creating new column values and pandas.

This was not the most efficient, and in fact, it's going to probably take a moment, probably a minute or two to run.

But I'm gonna show it anyways because it always works.

And then we'll talk about it, will show a couple other examples that are going to be, like, way faster.

But anyway, uh, get eso it.

Let's do define, get, get you can't wage on.

This is a function is going to take two parameters is gonna take a year and by state, so recall minimum wages by state.

But our unemployment data's by county are voting data is also by county.

So, unfortunately, can't buy county.

The minimum wage actually does kind of differ.

So right away, we've got a slight violation here.

Um, but that's okay.

The show must go on.

And I'm really just trying to show you guys an example anyways, so whatever, but sure, that's that's a problem.

We just I don't have a minimum wage data by county, so I'm sure there's a file out there that exists.

Feel free to use that.

If you're If you're trying to make an actual report here, just take note that this isn't quite fair.

Uh, anyway, what we want to do, we'll do a try and then, uh, accept.

And if we hit an exception, it we're gonna turn n p dot nan also have we?

Do we We did important, right?

Because we used nan right using in here.

Wait, everybody just relax a second.

Where when did we We never imported numb pie.

Not on this one.

What?

Okay, well, you're gonna need numb pie.

Mmm.

I'm gonna twilight zone.

Okay.

Whatever Um, yeah.

Okay.

I'm still Okay.

Whatever.

Here we go.

Um, rude term act.

Men wait, because we don't wanna get the value.

Right.

So we're trying to get a minimum wage by year in state, so return Ackman wage.

Um, and we can't just say your state because that action Ackman wage is indexed by year.

So actually, Aaron do dot Lok, And then we grab the year, so dot Loke, what's that look year, Okay.

And then we'll actually, so it would be year, But no, we're gonna say that Lok year, Okay, as a reference to the index column, and then we want to get that specific.

Um ST.

So that's a function that will do that.

We can quickly test that function by doing something like this.

I get men wage 2012 and, uh, Colorado and, in fact, 2012 should be a string Cool.

So that's 8.33 Probably impossible to see, but hopefully on your screen.

You see it?

Okay, so now we've got that.

Now we're gonna map this function to a new column in pandas, and the way that we can do that is just like regular mapping function in python in general, so we can map a function to a list or array in python using map.

And this is like I said before, not going to be the most efficient way.

But this is a way that will always work.

Especially when you have things like functions with multiple parameters or you have a fun like you want it.

You're not just mapping like one value to another place.

Maybe you want to run some sort of true function or calculation or, um, algorithm.

Right.

You need some more advanced.

This is your This is your guy.

Okay, so this is how you're gonna do it.

So first of all, I'm gonna throw in time.

Just tow measure time here.

That is a python notebook.

Specific commands.

I don't do that if you're writing pie charm or something, I don't even know that might work in by term, but definitely not an idol in sublime text.

So we're gonna map a new column.

So it's bun emp county, and we're gonna say, Umm min, wage where men wage and that is going to be equal to, um I think I'm just gonna write map for now.

We have to, in case this is well, but we're gonna say we want a map, get men wage.

We're gonna map that function.

And then you just passed all the parameters of that function.

So year, what year are we gonna map on EMP County year, and then what's the state won't.

Same thing right on on EMP County, Um, state and actually, years A capital.

Why year?

Okay, so we map that.

Unfortunately, I don't think this will work.

I'm gonna run it just because I'm on fresh pandas and all that.

And I'm curious if it does, um, did that work?

Everybody relax.

A second on.

I have to check this now on on in County.

I didn't think that was gonna work because as I think like Python, I used to work and fight on two and then with python three, you actually had to do list map?

No way that worked.

No, you get uh, that's hilarious.

That's adorable.

Okay, so yeah, we do have to catch it.

So I'm planning to That actually worked.

You didn't get a map objects, so actually, we have to convert this to a list.

Oh, that was so exciting.

so fast.

Okay, so it actually be calculated.

Here we go.

And then whenever that is done, I just want it.

We could also see when it gets a number.

But I'm just gonna do a nympho county dot head here.

Cool.

Okay.

So again, all that's doing is we're just we're converting it to a list at the very end because that, you know, with with pandas, you you could just say a new column is a list of things, and it doesn't matter how you generated that list.

It doesn't have to be via python.

Well, it doesn't have to be via python initially, I guess, but doesn't have to be viable.

Pan does.

Okay, um but yet so this is going to relatively slow.

If you can do a dot map, that's great.

Otherwise, a dot apply or something like that.

There are some other options for us and all.

Kind of a show, an option, any way down the line.

But once that's done, cool, we are good to go.

So the first thing that I would do once that is done is I would just check to see unkempt county dots.

In fact, let's just do this on EMP County er rate and then men wage lips.

Did that capital W again.

So don't forget rate is the unemployment rate.

We price you to change the name at some point, but that's good.

Um, and then this is minimum wage.

And the question is, is there a relationship there?

Do they co relate together?

And then the next question would be pretty much the exact same thing, except to see, do they vary together, so co variance, Um, and we are still waiting on that calculation are Oh, my gosh, it's taking forever.

Um, so probably while we went on that we'll just load in the presidential data while we wait.

Um, And then if someone knows a quicker way to do multiple, like a quicker way to map a function with multiple parameters, so it can't just have one parameter, because I know, yes, there are quicker ways of it.

Just only had one, but what about when it has two parameters?

I don't know.

The cool thing is, though, that it just knows to map.

Um, you know, by row like it, it's still pretty magical, even though it takes a lot of time.

Okay, but anyway, we are done on, um, County had done.

And then here we can see.

Is there any correlation?

So it doesn't look like there is much of a true relationship, but it does look like they actually do very together pretty significantly.

Um, so that's interesting.

I guess that would probably suggest that there's other factors at play, and those other factors have a high correlation to these things.

But these things between each other don't.

That's kind of how I would read into it.

I could be totally wrong.

I'm just here to teach you guys pandas not toe, actually, uh, find true answers.

It's Anyway, let's load in the presidential data now.

So say Prez 16 equals P d dot read C S V.

And this would be data sets slash prez 16 results stopped C s V now, uh, prez 16 dot head.

And when I get here is, uh, County Phipps.

I'm not sure what Phipps is.

Candidate s t, which I believe is state, but I don't know for sure what S T stands for.

Ah, percent report.

That's probably how many police is reported.

I don't know votes.

That's I guess?

Total.

Oh, this is how many votes for that person.

This is how many votes in total were made?

Yeah, and then this is what percentage for what?

Which candidate.

So as you can see, so for each, um, for each county in each candidate, we get all this data.

But to keep things simple, let's just do Donald Trump.

And then the value we're gonna track is the percent.

So for each county that we find, our question will be So how many people in that county voted for Donald Trump?

And we'd like Thio check correlation co variance between the unemployment rate in the minimum wage in their relationship with the people voting for Trump.

Okay, so let's do that.

So how do we do that?

Because now we actually need to combine all this data on county and state and then share the percentage vote.

So that's that Sounds like a really complex thing that's really hard to jumble in your head toe.

How you might do that, but it's not too too too hard.

Curiously, um, I think I think I'll just continue on this, but, um okay, so So what?

We want to d'oh now is Let's just grab, because our our, um, unemployment data on EMP on hemp.

I'm sorry.

On a planet county, we're gonna call it 2015 in the second Uneme County.

Uh, let's just do a lend of a nem County, right?

That's a very big data frame.

So I think what I'd like to do is just get the latest unemployment county data, which I believe is February 2015.

And, um, yeah, so we could make that much smaller.