Practical Charts for Finance Teams to Turn Data Visuals into Strategic Tools with Nick Desbarats

People tend to think of charts as visual representations of data, but I actually don't really think of them that way.

Affects for doing a job, and this is something, but especially for people with finance backgrounds, this is a big kind of leap for them.

Like when they sit down to create a new chart, typically they'll ask themselves, what is the best way to visualize this data?

I don't think that's a good starting point because the reality is that for any kind of data set, there's a literally infinite number of ways you could visualize it.

Welcome to another episode of FPNA on Locked.

Are you tired of being seen as just a spreadsheet person while others get a seat at the table?

Well, then welcome to FPNA on Locked, where finance meets strategy.

I'm your host, Paul Barnhurst.

Many of you know me as the FPNA guy.

Each week we bring you conversations and practical advice from thought leaders, industry experts, and practitioners who are reshaping the role of FPNA in today's business world.

Together, we'll uncover the strategies and experiences that separate good FPNA professionals from great ones, helping you elevate your career and drive strategic impact.

Speaking of strategic impact, our title sponsor for FPNA on Locked is campfire.

The ERP that's helping modern finance teams close fast and scale faster.

Today's guest is Nick Deborah.

I'm super excited to have him on the show.

So Nick, welcome to the show.

The light to be here looking forward to it.

Yeah, so let me give a little bit of background about Nick.

But first I'll tell how I first learned about Nick.

So I shared a post working on my data viz course.

And one of the people I've had on the show that many of you have listened to.

And I always recommend his book as Brent Dykes, effective data storytelling.

Comedy goes, someone else you should be following.

Someone else you should be looking at that doesn't really great work here as Nick.

And then you went ahead and commented and you shared a post on YouTube.

I believe it was about when to use tables versus graphs is where that started.

And so we got in touch and you agreed to come on the show.

So I feel honored to have you.

Well, feelings mutual.

It's a great job and listening to some episodes.

Well, thank you.

I appreciate that.

So let me give a little bit about Nick's background.

And then we're going to jump into data viz today.

We're going to go deep on that topic.

So as an independent educator and bestselling author, Nick has taught data visualization

and information dashboard design to thousands of professionals in over a dozen countries

at organizations such as NASA, Bloomberg, Visa, the United Nations, Shopify, the internal

revenue service, and the central bank of Tanzania.

Nick is the author of the Amazon number one new release practical charts.

I have it here thanks to Nick.

He sent me a copy and the upcoming practical dashboard books.

And he regularly contributes articles to the journal of the data visualization society.

Nightingale that are among that publication's most widely read and shared.

He also regularly delivers main stage talks at conferences such as the Tableau conference,

TD, WI world conference, SaaS explorers, data innovation summit and others and has lectured

at Yale Columbia, the University of Toronto and the Victoria University of Wellington in New Zealand.

So love the background one last thing to share on his background.

He was also and I really found this fascinating.

The first and only educator to be authorized by Steve and few to deliver his foundational

data visualization and dashboard design workshops.

And I just have to say probably the first book I bought on data visualization

was show me the numbers.

So I a huge fan of Stephen.

You got some coffees right up there there.

There we go.

Yeah, yeah, I have it.

I have it right over here.

I could grab it.

So I get it.

So you taught that from 2014 until launching your own workshops in 2019.

Prior to that, he held senior executive positions at several software companies.

And it was a co founder, a bit flash, which raised over 20 million of venture financing

and was sold to open text corporation in 2012.

Nick was granted a United States patent and the decision support filled love the

background is a great background.

You have their Nick.

Yeah, when people asked me how I got to where I am.

My one word answer is like circuitously indirectly kind of all over the place.

But yeah, it's this is now the best job I've ever had.

So can't complain love it.

You know, there's nothing better than enjoying what you do.

My training partner I've had on the show several times wrong on tarot.

He wrote a book called Love Mondays.

You know, the whole idea to help people find more fulfillment in their job

because so many people we've all seen it's like, is it Friday yet?

I hate my job.

I went when when can I leave?

Is it I could leave up five votes for 50?

Maybe if I just punch out earlier, you know, whatever the kind of things are.

So I'm glad you love what you're doing.

Absolutely.

Yeah, wouldn't have it any other way.

Good deal.

So I modified her first question.

Usually I ask what makes for grade F PNA, but I wanted to ask you a different question

since we're focusing on data is what makes for a great data visualization person?

What do they need or what do you see in those that are great at it?

Unfortunately, the most common answer typically revolves around software expertise.

People think, oh, you know, if you want to be really good at making charts,

you should be really good to use Excel or Tableau or some other data visualization product.

But I think that's only actually actually a relatively small part of the skills that are needed.

And actually, maybe if I can share my screen, I have kind of a sort of a list that I often use

and talks and presentations about.

So if you're listening to this on audio, we encourage you to go to YouTube.

Yes.

Yeah, I mean, I'll kind of rhyme off the list that I'm showing on the screen right now.

But basically, like I said, most people tend to assume that, you know,

good software expertise makes you great at creating charts.

So that's kind of like saying, you know, if you want to be a great writer,

become an expert user of Microsoft Word, you know, it's like, no, you know,

there's like, yeah, you need to know how to use a word processor.

But there's all sorts of other skills on top of that.

And so the list that I'm showing on the screen right now is showing things like data handling,

expertise, basic statistical knowledge, situation on all of the data is fundamentals.

And those are skills that I consider are necessary to create any kind of chart.

So regardless of what kind of chart you're creating, I think you need to know those.

And then there are other skills like storytelling, graphic design, coding,

advanced data visualization, software expertise, advanced statistical knowledge.

And these are skills that are required to create certain types of charts.

Like if you're creating, for example, data art or an infographic or a very technical or scientific chart,

these are additional skills that you would need.

And so really it's, you know, just focusing on, you know, software expertise.

Really, I think is, it's like, it's only a tiny fraction, in fact, of what you need to know.

And unfortunately, in a lot of organizations, they don't understand this.

And chart creation is often kind of sort of shovel off to more junior people

who perhaps haven't mastered a lot of these skills, especially things like situational knowledge.

Like if you're creating charts about insurance, then you better know something about insurance, right?

How, how that domain actually works.

And I see too many people basically saying, oh, you know, I'm a data visualization expert.

I can create charts about anything. It's like, no, you can't, you know, like you have to,

it's like writing about a topic, right?

If you don't know a lot about it, then your charts are not going to be very good.

I agree. And I like how you said, you know, the software is a small part of it.

That's what I say with financial modeling.

So if I'm just an expert in Excel, I can be great at financial modeling.

No, you need to understand accounting.

You need to understand business, how to validate assumptions.

It's very similar to what you're sharing here.

And I think so often people want it to just be the software because that's easy to measure.

And that's a technical skill that you can just go out and learn.

But it's not where the great, I think great charts come from that.

That come from understanding so much more science.

There's a little bit of art in that judgment.

Exactly.

I mean, that I put under kind of database fundamentals.

And so that's things like, you know, I have to choose chart types and how to,

you know, should I include zero in the scale, you know,

should I use a sequential or divergent color palette?

And that's really important.

And so it is in the list here.

But again, you know, it's just one piece, right?

There's amongst many that are needed, I think.

I agree.

It's an important piece because that's what I teach.

But I would never say it's the only piece.

That's for sure.

Yeah, I think it's a great point.

I know we'll come back to some other lists throughout that you're going to share

and a few other things.

But I'd love to have you tell your story of how you got started.

You were the only educator authorized for some of the workshops

that Stephen Fue offered.

How did you go from an executive out of software company to teaching

Stephen Fue's data visualization?

That's quite the route.

Yeah, yeah, I know it's like I said, indirect.

But I guess it probably really started originally.

It was a software developer.

And so I've always had an interest in data and coding and whatnot.

But then, you know, probably 20 years ago or something like that,

I just started to develop kind of the sign line interest

and cognitive psychology, the psychology perception,

you know, how we process information, how we make decisions.

And so I just started inhaling a lot of books on that topic.

I read, you know, over a hundred books, I think.

But I always considered it to be my kind of my sideline interest.

It wasn't my day job.

But then in 2013, I attended a workshop by Steve,

by Stephen Fue.

And if anybody listening has ever been to one of his workshops,

it's a magical experience.

I mean, he's just an incredible educator.

He's retired now, but, and, you know, just blew my mind

because it was kind of the intersection of these two areas of interest,

you know, data and technology and psychology, right?

Because in fact, most of what he talked about in his courses

really had more to do with psychology than to do with data

or technology.

And so, you know, I was fascinated.

We've kind of stayed in touch for a little while.

I got laid off company shrank by a lot.

And then that was my first phone call was asking Steve,

like, has anybody ever really approached you

about teaching his workshops?

And people had, but they tended to have very kind of data,

heavy backgrounds, not really kind of on the psychology side.

But because we had been in touch for, you know,

time after the workshop and he knew that I had kind of a strong

kind of base in that at that point.

He agreed, and so spend about seven months, pretty intense,

getting up to speed, sending me a pile of books to read.

And then, yeah, so I basically sort of taught his courses

for about six years, as you mentioned.

He retired in 2019 and encouraged me to sort of, you know,

launch my own courses, my own books, which I did straight

into the pandemic, of course.

But it turned out okay, because it turns out that people are

actually now, now they're okay learning online.

And so right now, still probably about 60 or 70%

of the workshops I deliver are still online,

even though I offer in person.

But of course, everybody hired so many people now that trying

to get everybody in the same city as sort of a big deal.

So yeah, kind of like I said, an indirect route to, to where

I am today.

Thank you for sharing that.

And so I got to ask, since I know you know,

a lot of the cognitive psychology, I think there's some great

books out there in that area.

Do you have a favorite or one or two you might recommend

to people on that subject?

I mean, I guess it kind of depends a bit on what you're

interested in, but like also be like how new are you to it?

Like, I mean, the kind of the Bible is thinking fast and slow

by Donald Countyman, right?

I figured that's the one you were going to say.

And, but it's, to be honest, it's, it reads a bit kind of

text bookie.

And so there are other books.

It's one of the first ones that I read.

Maybe we can put it in the show notes.

But I made the stick also incredibly useful book to just

basically about how to communicate information in a way

that stays with people.

Those are probably at a top of the list, I would think.

Okay, I haven't read made the stick.

I've read stories that stick.

That's a good book, different little bit different, but

really good about just how you, you know, the psychology

behind making stories stick.

Yeah.

Yeah.

Cool.

Thank you and appreciate that.

And I do tend to agree with you thinking fast

slow is a little bit like a text book.

It's good, but it does read a little heavy at time.

Yeah, I mean, in terms of like pop kind of more popular

books drive by Dan Pink or any of Dan Pink's books are also

great kind of first books to kind of start to get familiar

with some of the research in cognitive science as well.

Yeah.

Makes sense.

He's a good one.

I have his selling is human.

Yes.

Yeah.

Also a fantastic book.

Yeah.

Yeah.

Another one that I would social psychologist, H Jonathan,

H-A-I-D-T, any of his books.

Excellent.

Yeah.

Perfect.

Well, thank you.

Those are some great resources for people because like you said,

there's the data to this side and there's the cognitive

psychology side of images and the storytelling and how you

make impactful charts.

One of my goals when I was writing my books and designing my

courses was to, you know, actually only include sort of the

minimum amount of theory that I thought people needed to to

know.

And when I went through it all, I realized that in fact that

amount was zero.

There is no theory in my books.

And because, you know, it's kind of the name, right?

Practical charts.

It's all practical basically, no theory because I didn't

realize that, you know, I mean, it's very interesting to know

about things like pretentive attributes of visual

perception or visual hierarchies.

But you don't actually need to know any of those concepts in

order to create what I call sort of good everyday charts, you

know, good enough.

Like it's not data art.

It's not, you know, super technical charts, at least the ones

that I focus on.

And actually, I don't think you need to know any theory for that.

I know it's kind of a controversial point.

And I've written a poster too about it.

Maybe we can link it in the show notes.

But yeah, it's very interesting.

And I encourage people to look into it.

But technically, if all you want to do is just create charts

that are going to go over well with your audience and that

nobody's going to laugh at or your boss is not going to shoot

you down, you don't actually need to know any of that.

Interesting.

I'm going to have to give that some thought.

I always felt it feel like it's helpful to know that, but I

could see where you're coming from.

So I'll have to read one of those articles.

Interesting.

So kind of speaking of the book, love the title practical

charts.

You already mentioned how the goal is to be practical.

But as I was reading through it, there's one thing that

jumped out to me that had to ask about.

And if the acknowledgment section of your book, you mentioned

social media and how many people on social media help you

with writing the book.

Here is talk, talk a little bit about that because that's

definitely in general, you're going to social media is not

the place I would go to for advice on charts.

Just because there's so much noise out there.

Yeah, well, it was more, I guess, not not so much.

I mean, I occasionally, yeah, I guess things that people are

posting, but more of the discussions that happen.

And so it was more around the things that I would post.

And then I would see how people would react to it.

Or questions, I sometimes I'd be going back and forth

speaking, oh, should I recommend this best practice or that one.

And so I just throw it out there and see what people said.

And nine times out of 10, there's stuff that came back where I was

like, oh, that's true.

Hadn't thought of that or that's an interesting way to kind

of maybe reframe the question.

There's also lots of garbage, of course, lots of comments

or replies.

I was like, no, that's really not helpful or whatever.

But invariably, I got a lot of useful material out of it.

To the point where I almost felt kind of guilty because I was

like, oh, that's a great idea.

Thanks goes in the book or the course.

And it happens during workshops all the time too.

Like I tell people at the beginning of my courses,

I update these all the time.

And it's often because I teach smart people.

I would never say I'm the smartest person in the room.

And sometimes people think of things.

They notice things that I hadn't thought of or that I had missed.

And this happens, I would say almost every workshop.

There's at least one thing, you know, even like a minor thing

where somebody points it out and I'm like, good point.

And then course gets a little better, right?

So now it's been iterated and iterated many times over the

course of the last, you know, five or six years.

And so I'm grateful, yeah, to all the people who have

basically answered me on social media or put out comments

during my workshops.

That makes sense.

I can get that and that is definitely true.

Often the best learning comes from others.

When you're teaching so often you're learning from what other

share like, oh, never thought about that.

Or I didn't know you could do that.

Or whatever it might be.

Yeah.

And just also learning kind of where people's knowledge,

you know, generally is like, what do they know already versus,

you know, what are they wondering about?

And so yeah, I consider every workshop that he just

basically marker research, right?

Get a better and better idea of what people find to be most

valuable or what they, they're like, yeah, I'm not so valuable

for me.

That's why I didn't end up writing the book until I'd been

teaching the course for about three years.

Because then I had a much better sense of what I could

cut because people didn't find it to be useful or they knew

it already.

And so really the goal was that it was like all meat, no fat,

like just stuff that most people probably didn't know already

about charts and talk about what motivated you to write the

book?

What was the kind of the impetus to push you to do it?

Well, really most of the heavy lifting happened with the

course, right?

Like I was, you know, because the book is, I mean, in essence,

just kind of almost like a transcript of me delivering my

course, right?

I mean, I did think it would be relatively easy to write

because that's what it was essentially, right?

It was just, you know, a transcript of me, you know, giving

the course turned out to be much harder.

And anyone who's written a book will tell you that, you know,

that process forces you to rethink everything and it totally

did.

I ended up going back and just ripping the course apart and putting

back together again afterwards, but probably the, you know,

kind of maybe the more pertinent question is, you know,

what prompted me to develop the course?

Like I was teaching Steve's courses before, which are excellent.

But I mean, I did notice some things that, you know,

when I was teaching those courses where it's like people want

to have more of an interest in certain topics.

And really, I felt that there were more opportunities

that to be more kind of specific.

So one of my goals in writing, well, designed in the course

and writing the book was to avoid the phrase, it depends

as at all costs, right?

There are a few times where I had to fall back on it and say,

I'm sorry, this is just too nuanced, it's too complex.

I'm going to fall back on, you know, it depends.

So you're going to need some experience and judgment.

But 95% of what's in my course and my, my books, anyone can

follow those guidelines, even if they have very little

experience because, you know, like should I use a line

chart or a bar chart?

Well, often, you know, the answers that I was seeing out

there pretty much well down to, well, it depends.

Sometimes it makes sense to use a bar chart, sometimes a line

chart, right?

And so I really went down all the rabbit holes and all of

this, the edge cases and everything to come up with in

in cage, this decision trees, which are very specific.

Right?

It's like, it's not just it depends.

It's what exactly does it depend on, you know, is the

story more about overall patterns of change over time, or

you're focusing more on individual periods, for example, do

have all the data, right?

Or are you missing a lot of data?

It does the data occur at regular intervals of time, or is

it the kind of data, like sales transactions where you

might have like five in one day and then nothing for two

weeks, right?

And so I basically isolated all of these factors and distilled

them into very specific guidelines that, like I said,

anyone can follow, even if they don't have a lot of chart

making experience.

And so that was really probably the main kind of motivation for

me to create the course in the book was that I felt that I

could be more helpful, especially to people, maybe you don't

have a lot of chart making experience by avoiding that dreaded

it depends phrase.

And I know it's one of my friends and colleagues, Andy

Cochreve, that's one of his favorite phrases, it depends, but

I kind of push back on that a bit and said, yes, that's

true, but it's also not helpful.

And so let's try and take them next step and find what does

it depend on, right?

Exactly what.

And then, but I mean, I'm also kind of careful to say like that

these are my, I call them guidelines not rules, because once

you do achieve a certain level of experience of expertise, you

might deviate from what I recommend.

And I say that at the end of the book, like that's cool, you

know, you totally can.

But the first kind of rule of knowing when to break the rules

is you got to kind of know what the rules are in the first place.

And then you can break them.

And so that's really where I see my, my role is in helping

you get up to speed on the rules, although I hate that word, the

guidelines, because a lot of my audience, they're not passionate

about data visualization, you know, this chart is just something

they have to do as part of their job.

And they're like, I just want to create competent charts.

And I want to learn that as fast as I can.

Great.

Okay.

I can do that.

I just want to be able to do it for my job and move on.

This is not a career for me.

Yeah.

I mean, some are right.

I do have people who are very passionate about it and you attend

the courses as well.

But I'd say probably two thirds or something are like, great.

You know, check.

Now I can create competent charts, you know, next, the average

person out there wants to be able to do competent charts.

I have to do them.

There are those that are completely passionate and love it and

want to go deep.

I hear you.

And so it kind of leads me to the next question going back a little

bit to the book.

But in the book, you mentioned the data visualization has its

own spelling and vocabulary.

Can you explain what you mean by that?

Labyrinth on that a little bit.

In a lot of ways, I think learning how to create a factor of

charts is almost like learning kind of a new language.

And in fact, there are a lot of similarities.

I think between, you know, language and charts more than a lot

of people realize, like, you know, a lot of people tend to think

of charts as very kind of technical.

But I mean charts are for only for people, right?

Charts serve no purpose for computers.

Like computers don't do not need charts and graphs, right?

And so in a lot of ways, it's kind of like charts are almost

like essays in a lot of ways.

And there are kind of sort of specific guidelines as I was

just mentioning that you can learn just like when you're

learning a language, like, you know, like you can imagine if

somebody was learning English, let's say, and they were trying

to learn when to use the three versions of their, right?

E-H-E-I-R-T-H-Wire.

And imagine if you had an instructor who said, well, you

know, it's just something that you'll sort of get a few

or four over time and with experience, you'll eventually

kind of, you know, figure out which one to use.

Whereas if you have a good instructor, let's say, no, no,

okay, here's the situation in which you would use T-H-E-I-R,

right?

It's when it's to indicate, you know, possession or whatever.

Here's a situation where you use T-H-E-Y,

right?

And when you could substitute with like, they are.

And so you can learn it, right?

It's not this sort of touchy feeling as a lot of people

present it to be.

That makes a lot of sense to me.

So it's kind of that idea of, look, there are, I use that

term loosely rules that you can learn.

I know you mentioned guidelines, but there are things that

can really help guide you in building more efficient

effective charts.

Yeah.

And without relying on having just years of experience

in judgment and intuition, it's like, I don't know.

I mean, that's required in certain cases.

But there's opportunity to be much more concrete about it,

I think.

And now a brief message from our sponsor.

Campfire build a best-in-class AI native ERP.

Designed for modern finance and accounting teams.

With Campfire, you can automate revenue,

streamline your clothes, and centralize accounting

all in one unified platform.

And now they've hosted Finance Forward,

a first of its kind, AI Summit in San Francisco.

They brought together the sharpest minds

in finance and operations, all focused on how to actually

use AI to move the needle.

You don't even have to travel to learn the tactical skills

to put your team ahead.

You can stream the whole thing on demand totally free

anytime after the event.

So if you're ready to get tactical about AI and finance,

head to campfire.ai and register for the on-demand content

for free.

That's Campfire, where finance teams go to think forward.

So let's get into a little bit more of visualization

kind of for the finance FPNA audience.

So the first question I'd like to ask is I know you do

a lot of training for finance and FPNA people.

Obviously, we build a lot of visuals.

What are the biggest or most common mistakes you see

from kind of FPNA finance people in general?

Well, I mean, I think probably even more broadly.

So yes, FPNA people, but just generally,

people tend to think of charts as visual representations

of data, but I actually don't really think of them that way.

I call them graphics for doing a job.

And this is something, but especially for people

with finance backgrounds, this is a big kind of leap for them.

When they sit down to create a new chart,

typically they'll ask themselves,

what is the best way to visualize this data?

I don't think that's a good starting point

because the reality is that for any kind of data set,

there's a literally infinite number of ways

you could visualize it.

I think a better starting point,

this is what I teach in my course in my workshops,

is, do I know why I'm creating this chart?

Am I trying to persuade the audience to do something?

Am I trying to explain something to them?

Am I just trying to make them aware of a problem

or an opportunity?

And if so, exactly what is that problem or opportunity?

Or am I trying to answer a question?

If so, what exactly is that question?

To me, that's the starting point.

It kind of has to be because if you don't know

why you're creating a chart,

then a lot of your design choices

are actually going to be random, right?

Things like choosing chart types and color palettes.

Those depend actually mostly, not entirely,

but mostly on the job of the chart.

Yes, the nature of the data is a factor, right?

How many values do I have?

Are these time series values or the breakdown of a total?

Yes, those are considerations.

But they're almost secondary compared to the job of the chart.

And so really, the process of creating a chart

is kind of combining that information,

the job of the chart, the nature of the data,

and then coming up with a chart

that has as high a chance as possible of doing its job.

And so if we were, for example, creating a chart

to persuade people to donate to our charity,

what ultimately matters is how many people saw that chart

and made a donation, right?

And so things like, well, you know,

how quickly can they interpret it

or how precisely can they perceive the values?

It's like, yeah, those matter,

but I call those kind of the secondary attributes of a chart.

The primary attribute is, did it do its job?

Regardless of like how much cognitive effort it took or whatever.

And these are often things that are measured in labs

when people are studying data visualization.

They measure like how much time it takes people to process

or how much information they could recall

or how precisely they can perceive values.

But to me, kind of missing the point.

I was like, yeah, you know, those matter,

but they only matter because those things tend to improve

the odds that a chart will actually accomplish the purpose

that we, you know, the reason why we decide

to create a chart in the first place.

And so for everybody,

this is a big mental leap, right?

To move away from what is the best way to visualize this data

towards how can we create a chart

that is going to do its job?

And oftentimes people, like,

particularly people with finance backgrounds,

but like I said, everybody in general,

it takes them halfway through my 14 hour,

you know, two full day in person for a half day online course

until they really start internalizing that idea

and they stop asking what is the best way to visualize

this data?

You make a really good point, right?

The, you have to think about what are you trying

to accomplish with your audience

and does the chart help you accomplish that?

Yeah, oh, if you're, is there a decision

an outcome that you're hoping to get

from that visual you show?

Not always time series the best for this data.

It that depends on what you're trying to accomplish

with that data.

Yeah, well, be more kind of what kind of time series chart.

If that's the kind of data you're showing, right?

There's actually six or seven major chart types

for showing data or time.

People automatically tend to think,

oh, it should be a line chart, right?

Data or time?

It's like, well, it's sometimes, yeah.

But sometimes it should be bars or dots or a step chart, right?

Even in certain cases, a heat map,

if you have a lot of time series in a chart.

And so, but you can't really make any of those design choices

unless you know why you're creating a chart in the first place.

Now, out there in the real world, of course,

there are unfortunately situations

where we don't know, right?

We've just been told, hey, make a chart of this data.

And we're like, okay, why?

And the audience is like, well, I'm busy.

Just make a chart.

Now those are tough.

And I talk about those in my workshop.

I have this technique called spray and pray

where you basically create kind of a couple of different charts.

And you hope that one of them, you know,

kind of actually, you know, is what the audience

had in mind when they were asking for it.

But yeah, always coming back to the job, the job of the chart.

I like the spray and pray.

We've all done that before.

Yeah.

Well, a lot of people don't actually,

they'll actually try and create,

they'll put all their eggs in one basket, right?

And they'll, they'll, they don't know why the audience

wanted to see the data.

So they'll create a chart based on an assumption

about why they want to see it.

But like the odds of that assumption

is going to be correct or very, very low.

Because like I said, for any data set,

there's a literally infinite number of ways

that you could visualize it.

And what are the chances that you happen

to pick the one that is going to resonate

most with the audience.

If you don't even know why they wanted to see the data

in the first place, it's pretty low.

And this kind of explains why there's

a lot of dissatisfaction often amongst audience members.

Like charts get thrown back if you all the time.

You know, that's not what I wanted.

You know, what the hell is this?

You know, how come you didn't like, you know,

understand what I wanted?

And it's like, well, you didn't tell me what you wanted.

But real simple answer that you gave me no details.

Yeah, yeah.

And so it's, but it, but oftentimes people don't even ask, right?

So they get this requesting here.

I need a chart of this data.

And they're like, okay, boss.

It's like, no, you need to, you need to come back and say,

okay, no problem.

But I need to have a better understanding

of why you're asking me for it.

You know, like you said before,

is there a decision you're trying to make?

If so, what is that decision, like tell me?

Or is there, are you going to show this to a client?

And you're, because you're trying to persuade them to,

you know, you're trying to upsell them or whatever, right?

You got to know that in order to create an effective chart.

Yeah, you make some great points there.

So I'd love to ask, you know,

what advice would you offer for people to get

to improve data visualization?

Say they're, you know, earlier in their, early in their career,

a kind of finance people are just in general.

Where do you recommend they start?

Well, right here.

Not.

Yeah.

Oh, I know you have practical charts.

Here we go.

Read the book.

Yeah, yeah, no.

Well, that will get you basically one of the kind of nine or ten,

you know, sort of skills or areas that I was showing earlier, right?

But you need to, you need to know that, right?

The basic kind of spelling and vocabulary of database.

How do you choose chart types and color palettes for sure?

But yeah, it takes time because that's not the only thing

you need to know, you know, database fundamentals,

which is what I teach.

She also, of course, you have to know how to use pieces of software.

That's typically actually the easy part,

especially now with AI, right?

You can, if you're not sure how to do something in Excel or Google

Sheets or whatever, you can just ask and it'll tell you how to do it.

So typically, actually the easy part now.

But you also need a data handling expertise and basic statistical knowledge,

you know, understanding things like survivorship bias

and when to use means versus mediums.

Like, if you don't understand those things,

you're going to plant those, I call those statistical landmines in your charts.

You have to have that situational knowledge.

That's usually where the real heavy lifting is, you know,

especially if you're kind of new in an organization.

Well, you need to get up to speed on like,

what are people worried about in the organization,

what are the kinds of concerns, what are their objectives,

you know, what kinds of charts have they seen in the past

that they might be used to seeing?

I'm not saying that you need to replicate that,

but you should be aware of that.

If there are charts about like healthcare data,

then you need to start learning about healthcare, right?

And so, so yeah, like I said, you know,

I showed that list before just start trying to check off

as many of those as you can, essentially.

And it might take months or years.

I kind of have to laugh when you said, you know,

knowing what type of charts people like to see.

I had prepared what I thought were some really good graphs.

Almost everybody loved the way they showed some data

and the VP came back and said,

no, he just wants his pie charts.

You know, it's like 20 slices on them to show the two different years.

And I just kind of like, all right.

I'll take you about two seconds.

I don't know, I wasted all this time to build this nice chart.

But yeah, yeah, well, I mean,

well, I don't actually have an inherent objection to pie charts

unlike, you know, Steve, for example,

Steve, he never uses pie charts.

So actually, you know, I used to be a kind of a never pie

charger.

20 slices is a little lot.

Yeah, it was definitely not a great use case.

Yeah, yeah.

In my opinion, but yeah, yeah, yeah.

Yeah, well, always coming back to, you know,

the job of the chart, although, you know,

I have actually created like it was actually like a 25 slice pie chart.

But that was basically because the job of the chart

was to show that the market was very fractured.

So even things like that, you always got to come back to the job

of the chart.

I'm probably closer to Steven than you,

although I have come off that.

I'm not as strict as I used to be.

I'm like, generally, you know, if there's four or five and

that's what you want to do and it makes sense, go for it.

I don't like using them myself.

But as long as you use them effectively and you're not

abusing them, go for it.

That's kind of where I've come to.

And it is surprisingly tricky to actually know when it does

actually makes sense to use a pie chart.

You know, I have this whole decision tree for it.

And yeah, like and most people kind of get wrong.

Like they use a pie chart when when actually a bar chart would have

been a better choice or a tree map or a stack bar chart or

something like that because they're actually about eight

considerations there, you know, like do you actually have all

the parts is your story more about fractions of a toll

rather than comparing individual parts with one another.

Right.

All these things, you know, need to come into play.

But I can taste this to somebody we in within a few minutes

with a decision tree, of course.

No, it's a great way to think about it.

The decision tree and I could I could appreciate that.

So yeah, that's always a fun one is I like to say I put

pie charts and kind of that the art side of data visualization

in the sense of where people have opinions.

What I mean by that is there are some things that are lean

a little more toward rules and there are some that lean a

little more toward opinion and everybody has an opinion on

pie charts.

And now a brief message from our sponsor is your AI thinking

like a generalist or like a finance leader.

Here's the thing.

Most large language models weren't built for finance.

They're trained on everything from Reddit threads to recipe

blogs.

So when you plug them into your accounting workflows, the

results can be let's just say not balance sheet accurate.

That's why campfire the company pioneering AI native

ERP just made a huge move.

They've raised a 65 million series B co led by Excel and

Ribbit bringing their total to 100 million raised in just

12 weeks and they're putting that momentum to work with

LAM the large accounting model.

It's a proprietary AI trains specifically for finance

and accounting already hitting 95% plus accuracy on accounting

tasks.

That's redefining what ERP can do for modern finance teams.

Unicorn companies like post hog and replica are already

using campfire to automate their finance operations and

seeing real results.

You can see it for yourself sign up for a live demo of LAM

at campfire.ai.

That campfire.ai.

Yes.

Yeah.

Yeah.

Although, you know, when I published this very kind of detail

article about it in Nightingale, the journal of the Digital

Visualization Society.

And I haven't seen a lot of great counter arguments to it.

And I did manage to change a lot of minds.

And so I'm always a bit hesitant.

It's like, yes, of course, there is some art.

There is some science here.

But I think there's a lot more science than people tend to

assume that there is.

Actually, I can maybe show that decision.

Yeah.

That'd be great.

Yeah.

I'm just bringing it up.

As you can see, like, you know, pie charts are, they're in

there, right?

But there are, of course, other alternatives, other different

ways of showing the breakdown of a total.

And there are a lot of considerations, right?

In terms of what goes into this chart type choice.

And so when people ask me, like, when should I use a pie chart?

Well, this is true.

And this is true.

And this isn't.

And this is.

And this is.

And you should use a pie chart.

Like.

And so it's kind of unsurprising that people often don't use

them appropriately.

Because it's not as simple as like, oh, you should be showing

the breakdown of a total of these pie charts.

Like, well, no, all of these are breakdowns of tools, right?

There's other considerations that come into play.

However, I can teach this to somebody within a few minutes.

It doesn't require years of experience or a lot of art.

You know, it's mostly science.

It's a really good point.

And as far as the whole idea, that's mostly science, right?

It's a decision tree.

That's not art.

No.

No, there's no decision here and there.

And when you get done, you have, you know, you boil it down

to where you should probably go or the one or two.

Yeah.

I mean, option.

Yeah.

There's probably not a lot of subjectivity here.

Like, I was, I was asked people.

It's like, look, if you disagree with anything in here, then

tell me.

But once people step through it, 99 times out of 100,

like, yeah, I guess so.

You know, can't really see any, any problems with it.

There are some people who just say, no, no, like I'm,

I'm just ever getting these pie charts.

So I'm like, okay, obviously that's fine, you know,

but can you point out where this is wrong?

Right.

And I don't get good answers.

So yeah, you have the people that are just, I'm not going to

use a pie chart.

But yeah, my joke is that I always use and I kind of have fun

with one of the guys online because he's, you know, he's a fan of

pie charts.

I always joke pie is only good for two things.

Eating and playing Pac-Man.

Yeah.

Obviously, I do that jokingly.

I'm like, no, there are cases where it makes sense.

I just don't like using them.

That's more of a, I recognize that's a preference thing

versus a rule.

Yeah.

Well, I mean, that's fine.

It just, you just basically, you want to make sure that

you understand why.

Like, you know, what are the, you know, specific reasons?

And there are, for example, there are actually chart types

that I actually do not recommend using, you know,

like things like box plots, oddly enough.

I wrote a whole article about this a couple of years ago

and Nightingale as well.

And it turned out to actually be their second most read

article.

It really hit a nerve with people.

But nine times out of 10, because I got a ton of replies

on that, a huge response.

And most people are like, yeah, I think you're right.

I think we need to kind of retire this chart type.

Not because it's inherently bad, but because alternatives,

things like strip plots and distribution heat maps,

things like that are just easier to understand.

They're less subject to, you know,

misinterpretation.

And so, yeah, maybe we can put a link to that in the show notes

as well, my box plot article.

And there are actually about 10 other chart types that I actually,

at least for your everyday charts, I'm not talking data,

or anything like that, but for your everyday charts

for reports and presentations, I just don't recommend using,

not because they're bad, but because alternatives are always

better.

They're just easier to understand, less subject to misinterpretation,

things that connect its scatter plots, for example.

I don't actually recommend using those.

Because again, there are better alternatives, essentially.

It makes a lot of sense.

I have box plots or something I'm almost never using in the FPNA

or finance in the data and generally visualizing.

So I haven't even thought about that one.

But I'll take your word for it, right?

I've never thought I should be using this over that,

because it's just not a chart type I use in the work I'm doing.

Yeah, yeah.

And that's fine, although I would encourage people

to use distribution charts in general, things like strip plots,

where each value is a dot on a line kind of thing.

Those can actually be very useful.

And they're a bit kind of underused.

Like, oftentimes, if you're showing an average,

like a mean or a median, you really probably

should be showing a distribution chart of some type.

It just shouldn't be a box plot.

I'm excited to dig into that a little bit more,

but it makes sense on the surface.

I get that.

All right.

So here's one I want to ask you.

Almost all charting tools come with default charts.

I still remember early in my career going into Excel

and selecting the 3D default pie charts

with all kinds of colors and thinking it was cool.

And so how do you decide if you should use a default

or customize the chart?

Like, what advice would you give people?

Because I think that's where a lot of people start, right?

You just take what Excel gives you.

And maybe it's just me, but I don't think the defaults

and Excel are usually great charts

that are following best practice kind of guidelines.

Often not.

Often not.

It's gotten a lot better.

Like all database visualization products have gotten

considerably better in terms of the faults.

There's no these like crazy, super saturated colors

and heavy, you know, thick black grid lines and things like that.

Like they kind of got that perfect.

So part of it is yes, like the defaults are still not great

in a lot of cases, but even if they were,

the software will only kind of get you so far

because the software doesn't know the job of the chart, right?

It doesn't know why you're creating this chart in the first place.

You know, what question you're trying to answer,

what decision you're trying to support or whatever.

And so the software is always going to be limited in that respect.

Like it will only get you so far

and then you have to basically bring the expertise to the table.

Now, of course, this is going to beg the obvious question about AI.

You know, well, can we just dump some data into chat GBT or cloud

or whatever and have it produce an expert level chart?

And I try this all the time.

You know, every time as a new model comes out,

I kick the tires.

I try to use it for different kinds of data visualization tasks.

And it's still pretty hit and miss.

I find basically where what it'll get you is, you know,

well, first of all, you have to know what to ask it for.

And so that means you have to have data visualization expertise

to begin with, essentially.

You have to know things like I have to tell it what the job of the chart

actually is, any constraints, that kind of thing.

And so if you don't have data visualization expertise,

your initial prompt is probably not going to be very good.

But more importantly, you also might not be able to spot all the problems

with the result because they're almost always are, right?

You mean interval scale in your histogram,

but the intervals are not all the same size.

You know, if you don't know that that's kind of a no-no, right?

That's going to misrepresent the data.

Well, you're not going to spot the problem.

And so, you know, and the other thing is just in terms of kind of

brute efficiency, like most of the charts that I create

and I suspect that most people create are actually quite simple, right?

Mostly bar charts and line charts and pie charts and things like that.

And those, especially if you know how to use the tool well,

like Excel or Google Sheets or whatever you're using,

I can create those in like 20 seconds, you know, something like that.

Whereas if I have to type out a description, you know,

in full natural language, it just takes longer.

And so for 80 or 90% of charts, it's actually,

there's no gain there at all.

Like it's just slower.

And so, and then what remains is basically charts that are more complex, right?

Charts that would require many steps to create in Excel or Tableau or whatever you're using.

And so that's when I will say, okay, well, let's try AI for this one.

And it does take a while, right?

Because if it's a more complex chart, then it takes a while to actually explain what you want as well.

And then usually the result is like the first crack anyways is like kind of 80%.

It's like, okay, that's not bad.

It's not not exactly what I wanted.

And you know, if you have the expertise, you'll spot there's some issues.

You know, the color scale is off or whatever.

And then you go into this kind of fix it, loop with it.

Okay, change this, change that, move the legend up, whatever.

But that's where often things in my experience kind of break down.

Like you say, you know, move the legend up and goes, yeah, no problem.

And then it just doesn't move the legend up.

And then you, you, you have an unpublishable chart.

Or it introduces a bug that I can't recover from.

Because when it most AI tools create charts, what they're actually doing is writing JavaScript code

that calls a charting library like matplotlib or D3 or something like that.

And it introduces some bug at some point in iteration number six.

And that it just can't recover from.

And so in a lot of cases, you don't actually end up with a publishable result.

Or you end up with something that's like 80% of the way there.

But then I'll have to bring it into Photoshop or Excel or something and do the last 20% myself.

Because it just can't handle it.

So yeah, I don't think, I mean, I do use it.

Like, you know, for certain, you know, sort of relatively specific tasks.

But just assuming that, oh, you know, because AI is out there, now everybody is,

you don't need data visualization expertise.

You know, everybody's a database expert now.

Not now.

Like, and to be honest, I think with the kind of trajectory that I see large language models on.

I'm not an AI expert or whatever.

But I suspect that that's going to be the case for the foreseeable future.

Unless there's some like profound fundamental new kind of breakthrough or direction of research or something like that.

And I actually think that's the case for almost anything that requires expertise.

If you want a good legal contract, you should be a lawyer.

You have to know what to ask the AI for.

And you have to know how to evaluate what it gives back to you.

And the fact that, you know, it's, you know, this clause would be unenforceable in court.

If you're not a lawyer, you're not going to know that.

So if you want great charts, you kind of have to have that data visualization expertise.

Yeah, this gives us something I've been preaching around financial modeling.

We tested a bunch of these agents that will build models for you.

I've tested a lot of them now.

We did a whole podcast series and two conclusions.

I came to I said the average financial modeler will get a lot more benefit learning Excel and modeling better.

Then trying to get AI to build the model.

And what you're saying is kind of the same for data viz.

And I think it gets to this idea that AI as I put it is a magnifier.

If you know what you're doing, it will magnify that.

If you don't know what you're doing, it will magnify that.

It's just a question of time.

Yeah, yeah.

I think that's probably a good kind of framing in many domains, you know, probably most.

In fact, like, yeah, medicine, you know, if you want a good diagnosis, you should be a doctor, right?

You got to know what the relevant information is.

You need to provide to the AI and how to know when, you know, one of the diagnoses that shoots out is like, no, no, no.

That is definitely not it.

But if you're not a doctor, I would you know, right?

Yeah, it's like some guy.

I can't remember what he wanted to do, but it told him to take a certain medicine that was poisoning him.

And he ended up in the ER and almost died because he followed AI's advice.

Yeah.

Yeah.

Or people using his therapist like it's like, yeah.

Yeah, I would use it as my hate.

No, not a lot of people are.

Anyways, different topic.

Yeah.

Totally different subject.

But I think the message, whether, you know, and it applies to data visualization, look, AI can help you.

I think it's the bottom line.

I don't think any of us would disagree.

But if you want to know where to start, start with the fundamentals.

Yeah.

You need the domain expertise, just like with financial modeling, you know, as you're saying, or anything else that requires any kind of specialized expertise.

Yeah.

AI is just, I think your framing is good, right?

It's a magnifier.

And so if you don't have the expertise, it's just going to make you like, you know, a loaded gun, right?

It's going to point anywhere though, right?

It's not going to point in the right place.

Agreed.

So we're going to move on to a section I have.

I kind of call this the FPNA section.

And I'd love to get your thought for people who want to build, you know, great charts, dashboards.

What would you say is the number one technical skill they need?

I would say that, you know, probably chart type selection.

I call that a kind of a technical skill.

I know a lot of people wouldn't.

They would say it's more subjective.

It's more arts, more intuition.

But I think that is a technical skill that you can learn.

It's about, you know, having years of experience and a lot of, you know, judgment and intuition.

And so, you know, and there are many other kind of technical skills, like not just using chart types.

We're choosing color palettes and scale ranges and things like that.

But, you know, I would probably start there because choosing chart types is what causes a lot of not just charts,

but dashboards as well to flop.

And I see it everywhere.

You know, charts that, you know, it's a regular bar chart when it should have been a stack bar chart.

Or it's a stack bar chart when it should be in a cluster bar chart or whatever.

And so knowing, learning that kind of technical skill, chart type selection,

I think it's going to have a lot of bang for the buck.

Got it.

That is a good one.

I hadn't directly thought of that as a techno skill, but makes a lot of sense.

What about soft skill?

I guess, yeah, like maybe sort of empathy, not sympathy, but empathy in terms of really trying to get into the audience,

is head and understand, you know, especially if they've asked you for data.

So you can really understand like why, you know, are they are their concerns that they have objectives that they have, you know, kind of touch on this

a little bit before.

But that's a skill that unfortunately a lot of kind of data people sort of lack, right?

They want to live in their spreadsheets.

They want to live in their databases and their software.

And they don't want to go into the messy sales meetings and marketing meetings and, you know,

operational arguments and things like that.

And so one of the things that I really encourage people to do when they in my workshops, like, no, you've got to get out of the spreadsheet, get outside, you know, go to meetings, even if they're not,

they're not going to be talking about data.

They're just going to be talking about, you know, I don't know customer personas or something like that, go to the meeting, you know.

And so because if you don't get that, you know, kind of what's what's happening outside of the data and in your users heads essentially, then you're going to be of limited value to them.

You know, as you were saying this, preach the same thing in FPNA, get out of the spreadsheet, go sit in the meetings.

Funny how much you really learn with any job that you sometimes you think is technical and you can do all at your desk.

That empathy that's understanding is really critical to so many of those jobs.

So I love that you shared that because I talk about that all the time in different areas.

And I love that you think the same way, but just relating it to data visualization.

It's that reminder of get out from your desk.

Yeah. Yeah. And it would be the same if you were a coder or any other kind of quote unquote technical person.

I think the same vice would would apply. Yeah.

Agreed. It's a good reminder. All right. So we have a few questions just to get to know you a little bit better.

So the first one is, what do you like to do in your spare time? What do you like to do for fun when you're not charting?

I guess. Yeah. Nothing terribly exciting. Just running and there's one one unusual kind of hobby that I have, which is Israeli self defense.

I've been saying, yeah, that's a first on the show.

Well, Krav Maga, which is like, I'm not Israeli or anything, but yeah.

And I, yeah, I was anything about the political situation.

It's basically, it's kind of, it's a bit martial arts sort of, but it was, it was invented actually in the 1930s to because

fascist gangs were essentially attacking Jewish populations all over Europe.

And so they developed this sort of kind of fighting technique.

And I just kind of took it on a whim. I took a course years ago.

And I was like, this is actually super interested. I've never been interested in like martial arts or MMA or anything like that.

But this is, is actually, I do now see some parallels between kind of how I teach because people think, well, like fighting is all chaotic.

And it's like, you know, just with experience, you know, I guess you would learn how to defend yourself.

It's like, no, no, actually you can learn very specific, you know, techniques and ways of thinking that are actually going to allow you to survive, you know, much more likely.

Anyways, if you're actually kind of a kind of attacked.

And it's very practical because it's not martial arts itself defense, which means there's no rules.

You know, if somebody's attacking in the street, they can do anything, right?

They can do all sorts of things, even, you know, that, you know, you would never be allowed to do in an MMA, right?

You know, punch in the throat, you know, taking the groin, like, bite your ear or whatever.

The rules go out the window when it's survival.

Yeah. And so what do you do, though, when there are no rules? And so I just became quite fascinating.

And it's an amazing way to stay in shape. It just like, you know, it's everything.

Strength, flexibility, endurance, you know, the whole, the whole kind of shabang.

So yeah, that's my unusual hobby.

We'll do one more get to know you question here. I'm debating between the couple I have.

We'll go with this one. What's on your desk right now that would make people say that's so Nick.

Or could be on your bookcase as well.

Yeah. Well, the thing that's on my desk is actually what I was talking about.

It's this, it's a new book about a craft.

And it's actually written by, by my instructor.

He just, I just wrote a book. So I just started reading that, but he wouldn't say that.

So Nick, because most people actually aren't aware that this is my kind of my hobby.

So it's more typical with probably, yeah, like there's lots of books about, yeah, like this.

It's like psychology, psychology, psychology, psychology, psychology.

Yeah. So that would be more, more probably kind of maybe typical.

All right. We're going to do one more for fun because I just want to see what you say.

If you add a ban, I think I may know where you're going to go. One chart type from existence.

Which chart are you banning?

Yeah. Well, like I said, there's, there's about 10 that I don't recommend using.

And yeah, probably, probably number one would be, or not to his grams of box plots.

I figured that's what you were going to say, basically.

Yeah. And, and, and oddly enough, I think I might be well on my way to actually achieving that goal.

Because, like I said, the article that I wrote about this a few years ago went very viral.

And I know now it's been circulating, especially in like the research community where box plots are pretty ubiquitous.

And really forcing people to sort of like oftentimes they'll just use a box plot without even thinking about another chart type.

It's like, oh, I want to show distributions. I automatically use box plots.

That's what I do right now. It's not something I do very often. So I just default to a box plot.

Yeah. So, you know, but really, like, you know, I kind of lay out my reasoning in that article.

And I've had a few people, not a lot of people will come back say, well, you know, I still like box plots.

So I can continue using them. And I'm like, that's absolutely fine.

You know, you, you do you, but can you tell me why?

I don't get good answers.

And so I think maybe it could be that oddly enough, even though it wasn't really intentional,

that might be my legacy and the guy who killed box plots.

Well, we all got to have a legacy. Congratulations.

Yeah, I don't hate them. You know, it's just that there were better alternatives.

And I, you know, one of the big concerns I have though is that I have, I used to teach them, right?

And I just saw so many people get so confused by them.

And basically it made them feel stupid.

And so I was like, okay, if you absolutely have to use a more complicated chart type,

because a simpler chart type just can't say what you need to say.

Okay, sometimes you got to do that.

But in this case, there were simpler alternatives that could communicate exactly the same thing,

could do the same job. And so it was just making people feel stupid for nothing.

And so I was like, I think we should stop doing that.

Makes sense.

All right. So we've got a little, little longer than I planned,

but I think this has been a great conversation.

So just kind of wrap up if people want to learn more about you,

the courses and resources you offer.

What's the best way for them to get in touch or learn more?

But where you have available?

Yeah, my website is just practical reporting dot com all one word.

If you have people Google me, they'll find my LinkedIn page connect with me on LinkedIn.

I'm always happy to do that.

I'm the only person with my crazily spelled name.

I guess it'll be probably somewhere in the show notes or the video title or something.

So you can get the spelling from that.

And yeah, of course, always happy to connect with people.

There's a contact form on my practical reporting dot com website.

You can always reach out to me that way as well.

And I'll say I know you're more than willing to connect.

I know you gave me some advice and some of the stuff I was doing.

And I really appreciated and made some changes from that and

appreciate you carving out an hour of your time to chat today.

So thank you so much for joining us, Nick.

It was a great conversation.

Yeah, likewise.

Yeah, I really enjoyed it.

It was a great questions.

That's it for today's episode of F P and A unlocked.

If you enjoy F P and A unlocked, please take a moment to leave

five star rating and review.

It's the best way to support the F P and A guy

and help more F P and A professionals discover the show.

Remember, you can earn CPE credit for this episode by visiting earmarkcpe.com

downloading the app and completing the quiz.

If you need continuing education credits for the F PAC certification,

complete the quiz and reach out to me directly.

Thanks for listening.

I'm Paul Barnhurst, the F P and A guy.

And I'll see you next time.

Bye.