Skip to main content
Loading Explore guides…
Data Visualization in Business · Acedra
Home
Explore
Business
Data Visualization in Business
Sign in to save
Chat
If YouTube asks you to sign in, open the video on YouTube.
Watch on YouTube
Reload player
Transcript
73 segments
Copy
Hide video
0:00
When it comes to business communication, data visualization is really important. Of course, the structure, design and overall story of your presentation are all important too. But if you have any data in your presentation at all, how you visualize that data can really make or break the overall effectiveness of your message. Let's take this chart as an example. It shows tree sales over the last few years for a local tree nursery. But it's unclear what they're really trying to tell me.
0:26
compare that instead with the second chart, which shows the exact same data. But the message is much more clear amidst lackluster sales for most tree species, sales of Kwanzaa and cherry, this green line right here have grown steadily since 2020. This might not seem like much, but the difference between these two charts is really important. They both show the exact same set of data, but in the first chart, there's no clear message, whereas in the second chart,
0:51
the message is clear and easy to understand. Charts like that. First one are all over the business world, whereas charts like the second one are a little bit more rare. But the truth is, building a chart like the second one is not that difficult at all. All it takes is learning a few core principles really well in this crash course. I'm going to teach you data visualization skills found at places like McKinsey, Bain and BCG. You'll learn about choosing the right
1:14
chart type every single time and why it really matters for your message. Then I'll teach you my best tips for chart design. Then at the end, we'll do a complete exercise together. I'll go through every step from raw data to finish chart and I'll show you where you can download the same exercise file so you can follow right along with me. Plus, as a bonus, I'll show you where you can download this free cheat sheet that's going to help you choose the right chart every single time.
1:37
I promise you that by the end of this crash course, if you truly pay attention and try to apply the learnings, you'll be amazed at how much more effectively you can communicate Data. Let's get into it.
1:52
Everyone. Paul here from Analyst Academy, where we teach PowerPoint presentation design and data visualization training to consulting firms and finance teams and all sorts of other teams all over the world. If that sounds like something you're interested in, make sure you check out our courses over at the Analysts Academy dot com. And if you're watching this video, you might especially be interested in our data visualization for consulting.
2:12
Of course, it's kind of like this video, but way more detailed with hands on exercises, case studies, quizzes in a whole lot more. So make sure you check that out again that's the analysts Academy dot com. The first and most important part of effective data visualization is choosing the right chart type. To illustrate this, let's take a look at the example from the beginning of this video. Once again, this is showing tree sales for a local nursery.
2:35
So you've got your four different tree types here and it's broken down into different clusters based on year. This is what's called a clustered column chart. The takeaway for a chart like this might be something like Queen's on Cherry became the bestselling species in 2023, which you can clearly see right here. You're comparing each of these different species against each other. And in 2023 is when it first became the number one selling tree.
3:02
But then what if you put the data in a stacked column chart like this? Now the emphasis seems to be on the total of each of these columns. You're combining all the species together. So the focus gets put right here on the overall total. So a message for a chart like this might be something like overall tree sales have risen steadily since 2020, driven by Kwanza and cherry growth. Again, the focus is on the overall trend, but you can still sort of see the different changes from year to year.
3:30
Now put this data in a line chart and the message changes entirely here. You can kind of see this Kwanza and cherry data
3:39
increasing over time. So the message might be something like amidst lackluster sales for most tree species, sales of Kwanzaa and cherry have grown steadily since 2020. You can kind of see these other trees are flat to declining, but this one is growing. Again, this is all the same data for each one of these charts, but the message is changing ever so subtly. So in the first part of this crash course, I'm going to cover the five major chart
4:02
types, starting with the most reliable one, the bar chart. The primary purpose of a bar chart is to compare categories against each other, and this makes sense. You can easily compare the lengths of each of these lines. It's really easy to see which one is longest, second longest, and so on. It's easy to read, it's precise and it's incredibly reliable. Let's take a look at this example for McKinsey. title says Top CEOs identified disruptive technology, the economy
4:29
and geopolitics as the most important trends to act on in 2023. so what this is showing is survey results about the trends leading into 2023. end with a bar chart. It's really easy to see which one of these is longest, second longest, and so on. So what they've done is they've highlighted these top three bars right here.
4:49
These are the most important. It's clear, it's interesting, and it communicates the message really effectively. And that's what a bar chart is meant for. Again, you want to compare different categories against each other, in this case, these different trends. That way you can easily see which one is the highest or has the biggest value or what have you. And of course there's a few different variations to the bar chart. There's a clustered bar chart which kind of focuses on these subcategories.
5:11
Right here. You can pair this green, dark gray like gray against each other. There's the stacked bar chart, which focuses on the overall size of these different subcategories put together while still being able to show the relative differences of each of the subcategories. Then there's the 100% stock bar chart where all the bars are the same length, but the subcategories differ from category to category. So you can see kind of the relative changes.
5:37
when you're trying to decide between these different variations. The question you should be asking yourself is what's easiest to compare with a clustered bar chart? It's really easy to compare the different subcategories against each other versus a stacked bar chart. It's easy to compare the overall totals of all the subcategories put together. But the most important thing to remember with the bar chart is that it's easy for comparing categories against each other.
6:00
The next chart on our list is a column chart, and this one's really similar to a bar chart in that it's easy to compare these different categories against each other, but there are a few reasons why you'd want to use a column chart over a bar chart and vice versa. The primary reason is that column charts are good for change over time. That's because our brains naturally read these charts from left to right. Take a look at this example from Charter.
6:23
It shows the changes in Harvard's endowment fund by year, starting from 99, all the way up through 2022. And the natural tendency with this kind of chart is to read it from left to right. So a column chart makes sense, whereas a bar chart may be a little bit more difficult if it was from top to bottom. Notice how with this chart though, the focus is still on the individual columns and comparing them against each other. This says 4% decrease from 2021. It's comparing the total for 2021 to 2022.
6:55
And besides, with this error right here. So for this message, a column chart is a really good choice. Well, the column chart, you have the same variations that you have with a bar chart. You've got the clustered column chart, stacked column chart, and of course, the 100% stacked column chart. And again, the question you want to ask yourself is what's easiest to compare with each of these charts? Are you comparing subcategories, the category totals,
7:18
or maybe the subtle changes in categories over time? The next chapter in a cover is a pie chart, and this is by far the most controversial chart in the data visualization world. And there's good reason for that. But the truth is, the pie chart does have its place when it's used correctly. Unlike a bar or column chart, you don't want to use a pie chart to compare categories against each other. The primary purpose of a pie chart is to compare a value against the total
7:43
or compare multiple values against the total. So in practical terms, what this means is you're taking one slice and comparing it to the total of the entire pie. And the shape of the pie chart really lends to this because our brains kind of naturally associate circles with 100%. So it's not that hard for our brains to take a slice of this and guess roughly what the percentage of it is. So maybe this is 55%. For example, where people mostly go wrong is when they try to compare the size
8:13
of these different slices against each other, this one versus this one versus this one, because that's pretty hard to do. You'd be much better off with a bar chart. Let's take a look at this McKinsey example. The title says 63% of consumers pay for at least one loyalty program. And here you have a donut chart, which is basically like a pie chart. But with this kind of center taken out and what it's saying is 63% of consumers pay for at least one of these three loyalty programs or I guess
8:47
Amazon Prime, Amazon Prime and others, or just other programs. And what it's trying to emphasize with this title is this section right here.
8:57
In other words, it's saying 63% of people either have Amazon Prime, only Amazon Prime and other loyalty programs or just other loyalty programs. And this would be a little bit more difficult to show in a bar or a column chart or any other chart for that matter. So in this specific case, a pie chart or a donut chart works really well. but if the message were slightly different, let's say the percentage of non-subscribers is greater than the percentage of subscribers,
9:23
then this message wouldn't come through quite as clearly. Instead, you'd be better off with
9:31
a bar chart. The reason for that is it's really easy to compare the links of these different bars. So this would be non-subscribers and this would be subscribers.
9:41
Just like with the other chart. There are some variations to this chart. One we just saw, which is a donut chart, which is essentially the same as a pie chart, but with no holes. So you can put extra data in there. The only downside to a donut chart is it's harder to see the internal angles, which for some people can make things just a little bit difficult. But for the most part they're very commonly used and a good chart choice.
10:02
There are some other variations, like a sunburst chart or a tree map which follow the same idea as a pie chart, but look just a little bit differently. the next chart on our list is a line chart, and this is good for showing changes or trends or patterns over time. It's very similar to a column chart, but the difference is instead of focusing on each individual column, all the points are connected together. So you kind of focus on the data as a whole.
10:28
This helps emphasize things like downward trends or spikes, upward trends, anomalies in the data, things like that.
10:37
So for example, here we have a chart about the growth in the number of diversity, equity and inclusion roles since 2015 And if you take a look at the message here, it's talking about a meteoric rise. So a line chart is good for this because A, it's expressed over time, but then B, it's showing this like strong upward trend. Yes, you could still see that data in a column chart, but it gets emphasized just a little bit more in a line chart because you have these sloping lines.
11:03
The caution with line charts is you want to make sure that you're using them for data that has a natural order to it. So, for example, time has a natural order, but let's say you had different categories. If there's no natural order to this data, you can put it in any order you want
11:22
and you could be showing a trend that doesn't necessarily exist. So be careful to use this for data that's expressed over time or in some other natural order. There's a few different common variations to a line chart. One of them, of course, is to just put multiple lines on the chart. This is good for a few different things. One thing is emphasizing how one of the lines is different than the other, but sometimes you can also have a chart where you're
11:46
trying to show that all the lines kind of follow the same pattern.
11:50
another variation to the line chart is an area chart. Now we're going into a lot more detail on this chart in the course, but basically you can think of it as a column chart combined with a line chart, whereas a column chart kind of focuses on these as values in a line chart focuses on the overall trend in area. A chart does a little bit of both. Then of course you have a stacked area chart and a 100% stock or chart which again serve much of the same function as a 100% stacked column chart
12:20
where it focuses on how these how the make up of these different categories changes over time.
12:26
The next chart on our list is a scatterplot, and this is one of my favorite charts. The primary purpose of a scatterplot is to show the relationship between two variables. So imagine, for example, if this represented weight in this represented height generally speaking, the taller someone gets, the more they tend to weigh. And this is represented mathematically with the coefficient of determination or r squared. So with the scatterplot you can be showing a few different things.
12:55
You could be showing positive correlation, you could be showing negative correlation, or you could be showing no correlation at all, which in itself can be telling. Sometimes it's also good for highlighting outliers. So for example, if you had a dot way over here that you wanted to emphasize, or it's good for just showing where a specific dot falls on this plot, here's a pretty good example from BCG. What it's showing is price volatility
13:22
versus this VRT or penetration of variable renewables. And what it's trying to communicate with this chart actually is that they don't believe there's much of a correlation at all, which, like I said in itself can be telling The other powerful thing about scatter plots is you can add additional variables to make the information a little bit more rich. So, for example, if you wanted to change the size of each of the dots,
13:45
if you wanted them to represent something else, then voila, you have a bubble chart and again, you have that much more information to communicate to the audience. could also change the color of the dots, or you can even change the symbols themselves. However, I would really caution against this, unless you absolutely have to. The temptation with scatter plots is to introduce all these extra variables. It just makes it more confusing for your audience.
14:08
Focus on your main message and make sure that comes through clearly in your chart. so when it comes to effective data visualization, 90% of the time you're going to be able to use one of the five charts we just went over bar chart, column chart, pie, chart, line chart and scatterplot. in a recent video, I remade fancy charts into realistic slides. In every single case, I used one of these charts, and in other videos I followed the same pattern. So the key to a good data visualization is learning
14:34
when to use these charts and learning it really, really well. The best way to do that, of course, is with practice and repetition. But I promise you, if you can focus on this core set of skills, you'll get really, really good at data visualization. Once again, if you're interested in this sort of topic, make sure you check out our course over at the Analysts Guide to become. We go in much greater detail and really explain the nuances of each
14:54
of the different chart types, and it's complete with examples, exercises, quizzes, that sort of thing. So make sure you check it out. In addition to these core chart types, there are additional ways to visualize your data and we don't have time to get into all of them now. But one in particular that I want to talk about is tables. Tables can be especially good for two reasons. The first is when you need to show precise values. So, for example, here we have a basic table that shows sales
15:19
for different regions and a few additional metrics like share growth and the number of full time employees in the table would be really good here if I wanted to show specific values like the negative growth values. For example, for example, if my message was in the east region, we saw -1% growth year over year, then this would be a good choice because I'm calling out a specific data point. The other major reason is if you're trying to show multiple types of data in the same space
15:49
here, we're not just showing sales, but we're also showing the share growth and number of full time in place. And that would be pretty difficult to display in a chart. Is it possible? Yes, but it would just make things a little bit more confusing. There's a good example of this by BCG. So here's a chart from a presentation about the New York media sector. And what it's primarily showing is the percentage of New York jobs versus total jobs in each of these different years.
16:15
So, for example, in advertising, 12% of the jobs were in New York versus the rest of the U.S. And then in 2011, that jumped to 13%. So if these individual values themselves are important, then the table is a good choice here. But the primary reason to use a table here is because you're also showing this extra information. For example, the growth in New York City jobs. These are both percentages, but they're showing fundamentally different data.
16:43
Plus, you have these numbers down here where you're showing absolute values. These are salary numbers. Trying to put all of this information in a chart would be difficult and confusing. So here the table is a good choice. All right. Now that we've talked about the different ways to visualize your data, let's talk about design best practices. Now, there's a lot of different things I could talk about here, but I'm going to boil it down into two key ideas.
17:07
The first is to make sure your message is clear. Now, this sounds like a pretty basic thing, but you'll be surprised how many people really mess this up. You have an overall message that you want to communicate. Make sure that message comes through very clearly in the chart itself. And there's a few different ways you can do that. Let's revisit this chart from McKinsey. Remember, it's just a simple bar chart comparing these different trends against each other.
17:30
But they do a few different things to call attention to the message of the chart, which is that these three trends disruptive technology, economy and geopolitics are going to be the most important trends to act on in 2023. The first thing is they have a clear title, and this is the easiest and most important thing you can do for effective communication. Have a clear title that explicitly communicates the message you're trying to tell the audience.
17:58
The other thing they've done is they've highlighted the bars themselves, drawing attention to them as the most important trends. And of course, they've ranked these trends in order to make it that much easier. These are subtle things, but they have a huge impact in the overall clarity of your message. Figure out what your messages and find a way to communicate that clearly. their few of the things they could have done. They could have put a box around these bars, maybe an arrow.
18:22
Some additional text right here. all of these are ways you can draw attention to your message, which, again, is really important. Another really important principle of good chart design is removing the clutter from your chart. You'd be surprised how many things you can take away from your chart that really don't matter to the overall message or to the data itself. Here's an example from The Economist that highlights this really well. The title itself is really simple.
18:45
The subtitle An explanation also very simple and clear. But notice how there's not even any axes on this chart. There's no data labels right here to clutter the message. All it is are the categories,
18:58
labels in the bars themselves in notice. Also, have they only use one color for this chart? All of these things are a deliberate effort to help you focus on the actual data and not on anything else. We actually did a whole video on The Economist and why their charts are so effective. So if you're interested in stuff like that on Twitter, LinkedIn, in the description. So really good chart design comes down to two things Emphasize your message in the chart and remove the clutter from the chart.
19:26
And if you can do those two things and do them really well, your charts are going to look a whole lot better. All right, now that we've covered the basics, let's walk through an exercise together.
19:36
All right. So we're going to keep this really simple. Here's the data actually from the table example earlier. And all it shows is sales by region, the share of sales for each region, year over year growth and sales, FTE. if you want to follow along with this exercise yourself, include a download link in the description below or instead of following along, you can do the exercise yourself first and then come back and see the solution here on the video.
20:04
So let's say your manager says, hey, I want you to help me understand the relationship between overall sales for each region and sales FTE. All right. So we have our data here. We just need to figure out which chart type is going to be best for emphasizing that message going through each of the different chart types. First, you have a bar chart. A bar chart would be good for comparing these different categories against each other, but that's not necessarily what we're doing.
20:25
For example, we're not trying to say that sales in the north is greater than sales in the Southeast or that east central or excuse me, South Central has greater sales than all of them. A bar chart would be great for that, but that's not what we're doing. And likewise, a column chart. We're not comparing these against each other and we're not doing it over time. So come chart doesn't really work either. Same thing with the line chart.
20:48
We're not really trying to show the trend of any particular set of data. We're trying to show the relationship between these two sets of data. So Blanchard is out also. A pie chart would be good. For example, if we were trying to say that East and east Central account for a certain percentage of the overall sales. But that's not what we're trying to show. Again, we're trying to show the relationship between these two variables. So as you might have guessed, the scatterplot
21:12
is going to be a good choice here. All right, so let's put this data into a scatterplot and see what it shows us. By the way, for this exercise, I'm going to use a tool called Employer Charts. It's the chart building tool that we use and recommend over at Analyst Academy. It's really robust, lots of excellent functionality, and it makes chart building really, really simple and easy. It's built right into PowerPoint here. You just select the chart that you want.
21:35
You can customize it really simple and easily, which you're going to see here in a second. But again, it's something we recommend. You can get a free one month trial by following the link of the description. All right. So we're going to build the scatterplot. So we'll just go up here, go down to scatter, and you can just choose the size. We'll just keep it simple like this
21:54
and then just we just need to copy our data into the chart. Oops. So let's copy this. Double click on the chart brings up this Excel like window here.
22:12
We'll come back over to this one and copy sales after you
22:22
piece that in here. These two columns are for if you wanted to separate these into different groups. So for example, if you wanted, you know, northwest, east, east, central, north, all to be in a certain group in a certain color, if you wanted the rest of them to be a different color, that's what you would use this column for. And then this would be if you wanted to include an additional variable, turn it into a bubble chart. But we're not going to do either of those things
22:45
because we're going to keep it simple.
22:47
Doritos columns. And here we go. We already have the data that we're looking for, and you can see that there's a general sort of positive correlation here. So when you're trying to build an effective chart, remember, there's two things you want to focus on making sure your message comes through clearly and removing the clutter from your chart. So let's start with the message. So the first thing we're going to do is just add a simple title here
23:09
that's going to give us our biggest bang for the buck in terms of effort. Let's move this chart down a little bit.
23:17
All right. So let's just put a simple title in here. More sales have to e tend to lead to higher sales for each region.
23:30
Probably a better title out there, but we'll just stick with this one. It's pretty simple, but really the main focus of this is to emphasize the message in the chart. The other thing we could do here is we can add a trend like this is going to help emphasize that this is a positively correlated relationship between these two variables. Now, just to make sure we're being comprehensive, let's include a little subtitle here.
24:00
let's say sales FTE versus total sales USD a year here just to be comprehensive. So smaller title size. There we go. So now we have the basics of our chart in place. It's really just about removing clutter. And the good thing is with a chart like this, it's kind of already cluttered, so we can't really remove the axes here. I think that would take away from the effectiveness of this chart, it being a scatterplot, we could probably remove the labels, but they don't clutter the chart too much.
24:38
So I think it's okay. But one thing I do want to do is actually change the color of the labels to something a little less standout ish.
24:48
So the focus gets put on put on the dots themselves. And changing the color of these is really simple and easy.
24:56
and actually, I'm also going to increase the size of these just a little bit more. There you go. and with that we have a pretty solid chart. There's not a whole lot of stuff here, but it is a nice looking chart that clearly communicates the message without too much distraction, which is exactly what we want. if there's one thing I want you to take away from this course, it's this effective data visualization is not about learning every fancy chart type.
25:21
Yes, those things have their place, but for most people, the key to good data visualization is learning the basics really, really well and applying them. Before we go, I did promise you a free cheat sheet. And this is going to help you choose the right chart. It summarizes a lot of the stuff that we covered today, so could be a good handy resource for you. You can get that by going to the analysts academy dot com slash charts. Thanksgiving. Watching today's video,
25:44
let me know down in the comments if you have any questions about the content we covered or if there's something in particular you'd like to see covered in a future video.
Here to help you learn
Sign in to keep a private conversation with your course tutor.
Ask me anything about the material…
Sign in to chat