OK Cupid | EDA

Hi I'm Felix.

I'm here to present age trend data on the okay

QB deficit or given for the assignment.

Um So uh through this process we're able to examine

the data to a little bit of cleaning and certain

modifications um for the trends for the h train that

we're interested in exploring more and learning more about.

Um So they did they were looked at again is

the okay cubic data.

We could help from um location data one with latitude

and longitude that help us uh create a map and

then addresses um to specifically find out where people are

based of a much more detail.

But Overall interested just finding out that the location using

those two additional deficits.

Um When he came to data cleaning uh We picked

on variables we could do mean amputations on uh oh

and all the variables.

We could do a little bit, a little bit of

binning. Um And for us that was the h data.

Um As far as I've been is concerned and then

for amputation and that was going to be uh income.

We could have done same on other variables as well

such as race.

But that uh could gets very complicated pay fast.

So he's sort of keeping to the basics and since

we're interested to see the eggs trends uh What we

mean by age trend is uh as far as the

building is concerned is that we're able to group that

did and to millennials gen Z's get X's, baby boomers

and so on and so forth.

Um In order for us to be able to draw

certain insights from that.

Um And yeah, all right, those uh data cleaning and

now it's time to explore the data.

Um We're looking at the six distribution.

We find that there are more males uh than they

are females in our dataset.

And if uh the other line of sanction here was

looking at the X trains across whether it's gen z,

gen X and all that.

Um we want to assume that this will remain true

uh as we go through and we do do a

little bit of examination later on to confirm that or

to prove that assumption.

So diving into the age insights after the beginning, we

find out more millennials um are followed by gen X.

And gen Zs million years age I think stops around

38 and then gen z h starts from 18 To

about 23 ish and so on and so forth.

And we found out seniors were sort of an outlier

in this age category.

Uh because very few of them they were able to

find in the data set.

And then the next step was trying to find a

distribution uh of these different age groups across the U.

S. And it seems like more of them are concentrated

either on the east coast always cause whiskers having more

of them and particularly in California as we can see

from the image.

Um And then the next thing was to find out

what is the sex distribution across these different age groups.

And as we uh made an assumption at the beginning

that it was going to remain true across.

And we can after further examinations, we found out that

it is indeed true across um um these different groups.

And then um furthermore, we're trying to find out um

This is mostly to as we're going through our towards

the end, we're trying to find what we conclude from

from all this.

And if there are any recommendations we could make for

the platform using different signals in this case we pick

offspring offspring signal to tell us more about parenting people

interested in that or not.

And when you look across the groups between between the

two sexes, they were able to find um it seems

like the millennial group is much more interested in that

um than any other group.

And we do also see some follow up interest from

the uh, the gen x, which makes sense because uh

um sorry, the gen z, so Elliot in life.

Um So this is 18-28 as people approaching.

Uh let's see 25, is that maybe thinking about these

things? And then once you get into the millennial arena,

we see this being true, people getting more interesting uh

in having kids.

And also there was a good representation of people that

do not want that, but we wanted to see if

uh if that signals anything and if there is some

signal, what could be a possible recommendation or recommendation, we

can work on to develop a particular feature on the

Yeah from so our findings were um uh pretty clear

that analyze them to stem from the extreme ends when

it comes to age.

Uh seems like the seniors are allies in this case

and in this case.

And then when you look at the income as well

in the income variable, uh you find that these um

extremely rich people mostly after uh think 250 K.

Reported income.

And then when we run some numbers with the what

we're talking about before the parent signal whether people on

that or not um with people that are interested in

his, we found that the millennial group was much more

interesting that followed by the gen z's.

Um So in summary based on the trains are able

to see uh we can we can conclude that they're

more millennials on the Okay Cupid platform to start with

and with that um there's several things we can do

uh We feel larger group that was presented when we

look at the parenting signals, we found out this group

was particularly I had more interest than any other group

uh sensitive many of them um and um based off

that uh if we were to build a recommendation engine

we would want to um if we're to do much

as we want to match people who have this interest

and when you look at the group, assuming that we

wanted to recommend um um or someone else to to

to be recommended to partner with someone else uh with

regards to parenting.

We want to make sure that uh we stay closer

to the groups in this new case.

When we saw that the gen z's um they're more

in as far as being interested in, the printing is

concerned you have more jonesy's uh interested in that um

and Canadians being uh the leading group when it comes

to that.

So for example, if you to recommend um partners from

each of these groups, we want to um also be

aware of these signals, We don't want to pay someone

else who they say um maybe that is not something

else, the potential interest.

Um There is more um I think there will be

more success if we're to uh pear people along the

lines of uh common interest.

Thank you.

OK Cupid | EDA

Chapters

Audio TranscriptAuto-generated