If you're like me, you've been using

ChatGBT to do data analytics, but

there's some limitations that it has.

Like if the file is too big, it

sometimes just doesn't work and it

really struggles with good

visualizations, especially modifying

those visualizations. So, I've been

testing a bunch of different AI data

analyst tools. And I think I finally

found one that I like that solves those

problems. And what's great is you can

upload a CSV, but you can also just

connect data sources like Google

Analytics or your database. You can both

chat with your data or you can build

live dashboards within it. It's honestly

awesome. Hi, I'm Cody Schneider. I've

been running online businesses for like

10 plus and data analytics is a huge

part of running digital and let me show

you how to use this app. It's called

graph.com. Okay, so you can go and sign

up for free at graph.com. They give you

five free CSV uploads and you can do

unlimited messages. Then once you get

into the application, it's going to look

like this. You can hit upload CSV and

that CSV can be up to 512 megabytes. So

the file can be huge. It can be a lot of

data that you put into this. So I

already have data in the application.

It's actually this YouTube channel's

data. So I'm just going to open that up

and start a chat with the application.

So the first thing that it does is it

looks at the data set and it understands

what type of information is in those

columns and then it's going to

brainstorm some ideas, some insights

that it can show you. Once it's

brainstormed those ideas, then it starts

building the charts and it's going to

write some insights probably about those

charts. We'll see what it does though.

Awesome. So that just came back. Let's

look what it did. So, it made these

graphs. You can click into these graphs

and download these as a PNG. And let's

see what graphs it made. Engaged views

over time show fluctuations with peaks

in March and April 25th indicating

periods of higher user interaction. And

subscriber growth shows steady increases

early in the year with slowdown midyear.

Useful for tracking channel growth

momentum. So, this is great, but let's

ask some follow-up questions. So, the

question that came to mind is, do

certain videos do certain video lengths

get more views? So, I'm curious if

longer or shorter videos do better and

just total views that they get on the

channel. So, I'm holding command and

hitting enter. That's going to start the

thinking and planning process. Now, it's

building that chart in the back end. And

what's really great about this, in my

opinion, is you can ask these open-ended

questions like, do certain video links

get more? And it's going to understand

that logic and then write the code to

answer that question. And this data that

I've just returned back is actually

super insightful. So, let's read through

that real quick. From the chart, we can

observe that videos with moderate to

long duration, several minutes, tend to

get higher view counts, especially some

videos around 8 to 15 minute range, even

some long videos in the third. And then

it gives a suggestion. You might want to

explore further grouping videos into

length categories, short, medium, long.

So, I'm going to tell it to do that and

pursue that. So, click go. And then

while that's happening, I'm just going

to click into this. So, say I wanted to

take this graph and post it somewhere. I

can download that as a PNG or just copy

it to my clipboard and then paste it

into Slack or wherever else that I'm

trying to send it. You can also come in

and modify the data. So, say you wanted

to look at a specific period. I could be

the last 3 months. And the other thing

that you can do is modify the period for

the date. So, say I just wanted to look

at the last 3 months. I can hit apply

and it's going to update that chart

live. And so, now let's look at what it

did. So, I asked that follow-up question

that it suggested and it made these

categories. a long, medium, and short.

And it's average views by those

categories. So obvious winner here is

medium. And then it looked at long

versus average percentage viewed. And it

created this table automatically for me.

So category length, average percentage

views, and average watch duration,

average watch duration seconds. So it

generated both of these graphs for me. I

didn't even have to think about it. And

then it goes and it explains this data

in further detail, which is amazing. So

now I'm just curious. We're going to ask

it a really broad question to see it. It

says, "What can I do?" And now I'm just

curious. So I want to ask it a really

broad question. I'm going to say, "What

can I do to improve the revenue of this

channel?" And again, I'm hitting command

enter. It's going to and it's going to

start thinking. All right. And so from

that broad question that I asked that

generated these three different graph

types. And then it gave me the summary.

And the main key takeaway here is to

improve revenue, consider creating more

medium-length videos, which tend to have

higher view and engagement. Increasing

views engagement will likely boost ad

revenue. And this is pretty obvious if

you know how YouTube works, but the fact

that it can surmise that on its own is

pretty amazing. So now I'm going to show

you my favorite way to use the app. I

have multiple windows open in the

background and I do what I call vibe

analytics. I'm basically letting it work

for me in the background while I'm doing

something else. So I'm going to say list

five more insight ideas to explore. Hit

go. And then I'm going to go back to the

dashboard and I'm going to start looking

at another data set while that's working

in the background. So for this though,

I'm going to open up the graph

production database. And while that's

working in the background, I'm going to

go and start another chat with another

data source. But instead of using a

spreadsheet, I'm going to use the graph.

And now while that's working in the

background, I'm going to go and set up

another but I'm going to do it connected

to a database. So I'm going to use this

Google Analytics 4 connector. And I'm

going to say something like make me five

KPI charts for this website traffic. I'm

going to hit send on that message and

it's going to start generating that. So,

it's actually pulling against the

database that has all of the Google

Analytics 4 data. And that's just one of

the many data connectors that Graph has.

They have a ton in progress, too. Like

you can see Claio and Shopify, etc. And

while that's working, I'm going to go

back over to the YouTube data

visualization notebook. And I'm going to

ask it to analyze this first one. I find

it interesting. Let's do that. And then

let's also do the second one. And I'm

going to just say make both of these

graphs or make both these charts. I'll

just say do both of these. And now while

that's working, I'm going to go back to

this other one, see if it's completed.

And it hasn't yet. So I can spin up

another chat. And for this one, it's

just a retailed sales data set. And I'm

going to say list some insight ID and

hit go. And now we're back at the Google

Analytics chat. And it created these

three graphs. So shows new user growth

over time, total user growth over time,

total conversions over time, total

revenue over time, and then also user

engagement over time. And it includes in

that some insights about each of the

graphs. What are the things to

understand about what these graphs are

showing and transform that data into

information for you. Again, I can click

into any of these and I can modify it to

weekly, monthly and daily and also

modify the time. You can actually see

the code as well that it wrote to pull

this information. So if you are trying

to work with an external database is a

great way to do that. I'm going to say

based on this data, what else should we

explore? I'm going to hit go. I'm going

to jump back over to this other one

that's looking at the YouTube data

visualization. And it looks like I

messed up on this subscriber

demographic. So, I'm going to ask it, is

there demographic data? So, I'm going to

ask it, is there demographic data in the

data set? Let that run. And then, let's

explore this graph because it's kind of

fascinating. It's basically looking at

total views versus watch time versus

revenue earned and showing the outliers

here. And so, what you can see is the

watch time is actually pretty

consistent. Even if the total views is

higher, it's almost the exact same watch

time. And I want to modify this graph to

make it so it has multiple Y axis. So I

want to put views on the left Y axis and

then watch time and revenue on the right

Y-axis. So let's do that. Now you can

see here that it doesn't have

demographic data. So made that mistake,

but it clarifies for you based on that.

So I'm going to say something like

update the chart performance metrics by

topic you made to have two Y-axis. Put

the views on the right Y-axis. I'm going

to hit run. And then let's go back to

Google Analytics. And again, it creates

more data that it can explore. Looks

like it messed up on this chart. So,

let's hit retry. We'll have to

investigate why that's happening. But

this is interesting. So, it's total time

by mobile and by desktop. And that gives

me the idea to ask the question, does

mobile or desktop users spend more time

on average on the site? I'm going to hit

go on that. And I had a little trouble

with that. So, I'm just going to provide

the code from this. So, I'm just going

to go I'm going to take that SQL and

copy that. then go back to the editor

and put that in and say and say add a

second y-axis to this graph. Let's see

if it does that. All right, so it's

having trouble with multiple y-axis. So,

I'll leave the video there. Full

transparency, I'm Cody Schneider. I'm

the co-founder of Graph. We're building

this application in public and we would

love your feedback. Please go sign up

and it would really help me out if you

comment AI data analysts in the comments

below. It helps the YouTube algorithm.

And if you like content like this,

subscribe to my channel. We're going to

do tons more. Thanks for watching.

Cheers.