Common chart types
Assuming you have a clean and useful data set (beware, most data is not clean and needs some work before creating a chart), and assuming that you know what your data is about and which particular point you want to illuminate with your chart, you’ll have to decide on the chart type you want to use. Each chart type has strengths and weaknesses when it comes to showing and evaluating data. Let’s have a look at the most common ones.
Line charts are usually the best way to display changes in your data over time. The classic example here is population growth in a given area over a period of time, or the development of an exchange rate over time. Area charts are structurally similar to line charts, but the stress of the presentation is more on the ‘weight’ of the data than its ‘development’ over time.
Bar or Column charts are your best choice when it comes to comparing values from your data set. They really work in very similar ways, the main difference is that bar charts work better than column charts if your data includes negative numbers.
Stacked bars or columns work well if the data you are looking at is used to not just evaluate the values themselves (within e.g. each year) but also to compare each category of data over the whole series. This is a good way to compare growth between several categories.
Pie charts present the importance of items relative to the totality of the data. Even though your data might be proper numbers, a pie chart will display them in terms of a calculated percentage of the total included value. Pie charts can only hold ONE data set with all its categories. Tip: limit the number of categories, group the small ones together for easy reading. There is no point in having 10 categories smaller than 1% … make it into 1 category “Others” and specify in the documentation.
You might occasionally come across doughnut charts, which are really ring-shaped pie charts arranged around each other like an onion. My advice is to stay away from those as they tend to be confusing and difficult to interpret.
Radar charts are often underestimated. Similar to Pie charts, they work well when comparing relative importance to the totality of data, but they allow for MULTIPLE data sets with categories. Unlike the doughnut chart, the data is superimposed and a bit easier to interpret.
Scatter charts (also called ‘scatter plot’) will show the correlation of TWO variables, often to indicate there is a clear link between the two, e.g. air temperature and sales of ice cream (the hotter it is, the more likely it is for sales to go up).
Bubble chart are similar to scatter charts, but they can show the correlation of THREE variables. Let’s say you want to add the number of flavours sold to the Ice Cream scatter chart, this is the way to do it.
There are, of course, other types of charts, dealing with specific data of a more financial or scientific nature, like stock charts, or gantt diagrams, etc. For the sake of brevity and general usefulness, I’ll leave those for you to explore on your own.
How to choose the right chart?
In the first section, you have seen the main chart types and what they are used for. You can, of course, base your decision about which types you use on your own preference, or the audience you expect to see at your presentation. However, here are some tips to help you make your choice based on which story you want to tell with your data.
If you want to compare values, your best options are Line charts, Column/Bar charts, or Pies. Pie charts are limited to one series of categories, while Column and Bar charts allow for a small number of series, and Line charts are able to show a lot of series (although I would advise to limit the number of series for the sake of clarity).
If you want to clarify the composition of categories within a series, the Pie chart is probably your best friend for a single series, and usually you want to have multiple pie charts to show off multiple series. If, however, you want to cram several series into one chart, you might use stacked bars or columns.
If you want to visualise the distribution of data within the whole set, you might go with Scatter charts or Bubble charts. If you can simmer down your data to a level that allows for a Line chart, Column or Bar chart, that might be advisable for reasons of clarity and ease of understanding.
If you want to analyse a trend, stay away from flashy charts and stick to the basics: Line charts are universally understood, and Column or Bar charts are equally useful, if somewhat limited in certain cases. Columns charts allow for easy distinction of trend lines, should you choose to add those.
If the data gets more involved, and you want to highlight the relationship between multiple variables, you’ll have to get more creative and use Scatter charts or Line charts (two variables, depending on the relationship) or Bubble charts (three variables).
General tips when using charts
While the choice of chart is paramount to making your point and using the data to its full capacity, there are a couple of basic points to think about when creating your chart and making it look just so.
Making charts for the sake of charts is bad practice. Have you ever been subjected to a meeting with endless charts that didn’t convey any additional meaning? There is value in charts, of course, but only if they serve to clarify information in a way that goes beyond a table. Save your time and your audience’s time: if a table says it all, there is no need to make a chart.
Use consistent colours, formats, labels, etc. While having a bit of variation in charts sounds like a lovely idea, you may want to decide on a uniform colour scheme and assign certain content to a colour or shape. Your customer might just be happy to always see their own data in one colour or in a particular spot on the chart.
Limit data to the necessary and useful without omitting unflattering data! A common mistake with charts is the idea that “more is better”, inundating the audience with too much stuff on a chart. The idea is to make the numbers relatable! Stick to what you have to show on the chart and dismiss the rest. Separate into multiple charts (even on the same page?) to clarify the content.
When to use a Bar chart or a Column chart? This seems to come down to personal preference and aesthetic considerations: long labels look better on Bar charts, they are easier to read and the general direction things are going corresponds to how you would read a page of text: left to right, top to bottom. Column charts tend to work better for values, while Bar charts work better with percentages.
You may want to consider Stacked Bar or Stacked Column charts to display your data. They work similar to Pie charts as they calculate the percentage and stretch all bars to the same length. This helps compare the composition of multiple series of data.
Use only charts your audience is familiar with. If you are presenting to a group of financial advisers, a stock chart will be appreciated because they will know how to read the data. If, however, you need to appeal to a diverse (and often unknown) group of people, stick to the basics: bar, column, line and pie are your best friends!
Applying two separate axes to two drastically different sets of data can be good for comparing trends in two types of data with differing orders of magnitude (e.g. number of bills and value of bills per month).
You may have come across 3D charts. Admittedly, they look snazzy and can be useful to show data. They are also extremely complex to set up, look uniform and still show all the data so they can tell your tale. 3D charts are notorious for hiding smaller data points, and they are usually difficult to read, unless they only contain a relatively small number of data points.
In addition, there are a whole lot of additional options to make your charts stand out more: a range of trend lines, rounded lines, custom data markers, background images, calculated fields, etc. My general rule is “stay away from those, unless you need them to clarify something”. There is a danger of going overboard and wanting to beat the ‘other’ presentations by dazzling the audience. Keep in mind: the audience wants facts, most of all. If those facts are nicely dressed up, all the better, but that is not a prerequisite to get your point across.
As you can see, there is a lot to consider when creating charts, and that is without even touching on the technical aspects and the tricks and problems you might run into when preparing your data and transforming it into charts. My advice here is to make sure your data is obviously telling your story: that could be in the form of text, or tables or charts. If in doubt, choose the simpler one. It will save you time that can be put to better use.
Ask the ClutterMeister
Ideas to help clear away the mess in your homes and in your minds.
Feel free to share any of my posts, but please put in a backlink to the original blog post. Thank you.
Hi, my name is Tilo Flache. My mission: help clients declutter mind and space.
This blog contains pointers for your journey towards a happier living experience.