Bar graphs, line charts, scatter plots – which type of chart should you use? We’ll get into that, but before we get started, it will help to know the difference between different kinds of variables that make up these different charts.

### Discrete vs. Continuous Variables

**Discrete variables** can only be measured in exact numbers. For example, number of employees. You can’t have 12.5 employees (at least not without going to prison). That said, discrete variables aren’t necessarily measured only in whole numbers. For example, you can count the change in your pocket and know exactly how much money you have.

- # of cars sold
- Locations
- Sides of a coin

**Continuous variables** are more fluid, in fact, that’s a good way to think about it. You can have a gallon of milk, but you can also have 0.95 gallons of milk. Furthermore, even a “gallon of milk” is never *exactly* a gallon. Perhaps it’s 1.00034 gallons. Continuous variables are less “measurable,” if that makes sense.

- Square Footage
- Weight
- Volume

### Categorical Variables vs. Quantitative Variables

This one is a little easier. Think of categorical variables as qualities and quantitative variables as quantities. **Categorical variables** can be divided into neat little boxes and can’t be expressed with a number.

- Colors
- States
- Blood type

**Quantitative variables** measure something, and are expressed with a number.

- Distance
- Dollars
- Weight

### Bar Charts vs. Histograms: What’s The Difference?

To the uninitiated, bar graphs and histograms appear identical. For one, they both have bars. The difference is that **histograms compare two quantitative variables, while bar graphs compare a categorical variable with a quantitative variable**. To illustrate the differences, let’s say you run a sales department, and you want to visualize how your department is doing.

### Bar Charts

If you want to chart how much money each salesperson brings in, you need a bar graph. The y-axis represents the amount of money, and the x axis has a bar for each salesperson. Every salesperson represents a category.

When making a bar chart, you may want to arrange it horizontally, rather than vertically. For one, experts say bar charts are easier to read this way. It allows you to arrange rows in descending order. For example, you could put your top salesperson at the top of the chart. Horizontal charts also help you avoid giving people the wrong idea. When you see several columns next to each other, your mind can’t help but look for a pattern or trend, as if they were arranged by time. This can throw off your audience.

### Histograms

If you want to chart sales by day, this would call for a histogram. Time is a quantitative variable. A good way to tell a histogram is that every bar needs to go in a certain order. You wouldn’t put a bar for February after a bar for July.

Most of the time, histograms are best as columns going up and down, while bar graphs are best as bars going from side to side.

### Stack Charts

Stack charts are a kind of bar graph that shows more data within each bar. Think of every bar as its own pie chart. For example, the bar might show total sales, while each segment of the bar shows where those sales came from.

### Line Charts

Histograms are bold and easy to digest, but they can be a little vague. When you need more detail, consider using a line chart. Line charts reveal granular aspects of data and they’re a great way to show trends over time. They’re particularly well suited to volatile data. There’s a reason line charts are the go-to charts for showing stock prices. Stock prices are volatile. They go up, and they go down, sometimes several times a minute. Furthermore, each one of these crests and valleys can have huge implications. In a situation like this, a histogram wouldn’t make sense, even though it’s measuring the same variables.

### Pie Charts

Pie charts show proportion. They illustrate percentages of a whole. That said, pie charts are perhaps the most polarizing chart in the world of data visualization. Many designers argue that they’re hard to read, especially when they’re divided into several small segments. Pie charts make sense when there are only a few categories, and each category is pretty large. They start to get confusing when you’re dealing with more than a few categories, or when some of the categories are only slivers. Think of it this way, if you had an actual pie in front of you, it’d be easy to slice it into 8 pieces or so, but cutting it into 20 would be difficult. Furthermore, it’d be hard to cut a slice of pie that represents only one percent of the total.

### Scatter Plot

Scatter plots are designed to make sense of several, seemingly unrelated data points. For example, let’s say you want to chart time spent vs money spent on every sale. One sale may have happened quickly, but the salesperson had to fly across the country. Another sale may have taken a long time, but happened at little cost to the company. In between, you have another hundred points of data. You can’t simply connect all the dots, but what you can do is create a scatter plot. With scatter plots, you can get a general idea of what seemingly random data wants to tell you.

### Bubble Charts

Bubble charts show the relationship between three sets of data. Similar to a scatter plot, but a little neater, they’re useful for forming a cohesive chart from large amounts of data. When creating a bubble chart, note that the volume of data is represented by the area of the bubble, not the diameter.

### Heat Map

Heat maps show the density of data points. Like a scatter plot, this visualization offers a collage of data points that often show a general trend. Heat maps don’t always lend themselves to singular conclusions, but they’re great for showing a distribution of data.

### Map or Geo Charts

Map charts, or cartograms, overlay data on a map. A good example is when you watch election results come in. The use cases are pretty obvious, need to show data based on location? Map charts are a good choice.

### Pivot Tables

Pivot tables are essentially fancy spreadsheets. They don’t really count as visualizations. To the uninitiated, they look almost identical to a spreadsheet. The main difference is the ability to collapse and expand specific categories of data. It offers the completeness of a spreadsheet without all the scrolling around to make comparisons.

Hopefully this gives you a sense of what kinds of charts there are, and when you should use each one. To learn more about chart design, head to our Resources Page.

## No Comments