Today I’m writing about the importance of visualizing data points in a good way. I’m not sure how detailed I have to make this chapter, and as this is a passion of mine I will probably stop it at this level unless people want even more detailed examples. Tomorrow I’m continuing with visualizing processes, events, and concepts:
When you have a lot of data points, then get them into one or more graphs. Don’t try to show this type of information in tables or text. It will take people too long to read through them and to see some trends in the data. If you see some patterns you want to point out then add some trend lines or highlight the things you want to say, and if you have no idea how to interpret the data then say so and just show the raw data points.
• Clear – In some cases, your data will have microscopic noise, and you can easily follow what happens.
• Quite clear – Other times your data will have a bit of noise but it you can easily plot a trend line, and you can see what is happening
• No trend – The third case is where your data is just noise without any trends, or you have not yet spotted the trend. Show the graph if you think it is important, either to demonstrate that we don’t know what is happening or to ask people to try to find a pattern.
• Outliers – In some graphs, you will have outliers, either showing incredibly good or bad results. These are either extremely important or unimportant, and you can never know what they are until you try to find the reason. If your productivity at one point is twice as high, then it is either a glitch depending on wrong measurement or something else happening that impacted it. Or you have found something that will revolutionize your work. What I’m trying to say: don’t ignore these points just because they seem impossible.
Notice that I did not show any pie-charts or bar graphs. I’m not that fond of them as usually only show one point on a curve. It is rarely interesting to know the exact value of something; trends are much more helpful as they show the context as well as the current values.
Marcus Degerman 