Visualizing data with Chart Builder - SPSS Tutorial

- [Instructor] The best way to approach any analysis with a data set is by visualization. Take a look at what you have there. There's some really good reasons for that. Visualization to graphics are very in high information density, they bring a lot of information in a small amount of space. People, humans, are visual animals, and we're able to extract a lot of information from these. So it's a great way to start and see what you're dealing with. Over the years, SPSS has made some wonderful developments in the data visualization options that are available, and I want to walk you through some of these. Let's start by going here to Graph and let's choose Chart Builder. Chart Builder brings up this little dialog here that says you need to specify the measurement level. Now, I'm going to show you why they say that because it turns out that not all of the variables that are in the data set that I'm using, which is demo.save, which is one of the sample data sets, not everything's defined correctly, and it changes the way that some of these graphs work. But I'm going to just press OK for right now, 'cause most of the time, your data will be fine. The way the Chart Builder works is you drag in a template chart and then you start putting the variables in. When you come down here to the gallery, you have a choice of variations on bar charts, and line chart, and area, pie charts, scatter plots, 2D and 3D, histograms, a high-low chart, usually for finance, boxsplot, great for looking at outliers and then dual axis charts, which can be a little hard to read, so I generally avoid those. You can also construct things from scratch by going to the basic elements, getting to do IDs, and then adding things like the titles and the footnotes, but I find it by far easiest to go to the gallery, drag in one of the templates, and use it from there. Now, the easiest way to do this is going to be with a bar chart, so let's go to a bar and just drag this in. And then let's take a variable, say, for instance, level of education. Now, please note one thing about level of education, you can tell it because it's got a little ruler here that this is a scaled variable. It's done as an interval, or ratio level, quantitative variable. But let's drag it down here, just pop it into x-axis. And truthfully, I can just hit OK, right then and there, and that opens up the output, and you see that it produces a fair amount of text that goes into the command. Obviously, you don't need to do this, and you can disable that, but it lets you know what's going on, and here is our graph. Now, there's a few things to point about the graph. Number one, it's actually good-looking, it's got good colors, it doesn't have too many extraneous details, it's easy to see what's going on, that we've got more at level two than at level one, and so on. There is however, one problem, and that is because SPSS thinks this is a continuous or a scaled variable, it's also giving us the mean and the standard deviation, et cetera. Now, if we were asking about years of education, that would be appropriate, that's a ratio level, but if we go back to the data set, and let's go take a look at the data set, I'm just going to hit here on this red star to go to the data, when I click on that, here's the education variable. Let me just double-click on that. And you can see that we've got the values are, one, did not complete high school, two is high school, and so on. That's an ordinal variable, those are ordered categories, and we need to show that that's the case. So in fact, I'm going to come right here, and then come right to this Measure column where it says Scale, which is again, for quantitative, that's interval or ratio level variables. I'm going to click on that and just change it to Ordinal. And now, when I come back and look at Data View, you can see, for instance, that I now have the little one, two, three scale that indicate ordinal, let's go back to our recent commands. Just click here to go to recently used dialogues, go to Chart Builder. And you know, if this is your own data set, you can click this and tell it to go away. But now I'm going to do a level of education. But let's reset it. I'm going to drag in this one again, I'm going to get level of education, which is now defined as an ordinal variable. By the way, when you drag things into this window, it'll show you a little sample visualization that is not based on your actual data. It's a mock up that sort of ignores what the actual values are. So don't look at that and try to interpret that 'cause that's just showing you. Well, it will have bars and the bars will look something like this. Click OK. And now you see a difference. This is a proper bar chart. Now we have the five categories and it puts the labels on them, did not complete high school, high school degree, and so on. Whereas previously, we just had the values one, two, three, four, five. Also, you see how we have the mean and the standard deviation up here. We don't have that 'cause that's not appropriate for an ordinal variable. But this is exactly what we want. This is a quick and easy way to produce a graphic. And a bar chart is a fabulous graphic. It's probably one of the most informative things 'cause it's so simple, and it's so easy to read for what you're doing. Now, I do want you to know, there are a lot of other options with the Chart Builder in SPSS. So let me come and do a different variable. Let's choose this again, you can just hit OK and have that go away. I'm going to reset this. And let's do for instance, side by side boxplots. Now what I'm going to do here is I'm going to pick a variable like years at current employer. I'm going to put that on the y-axis 'cause that's my outcome variable, the thing that I'm trying to get a boxplot for. Boxplots are really good for showing outliers. And then I'm going to pick another, let's say, well, let's pick level of education. I'll put that right here in x-axis. And again, that's not what the actual distribution is going to look like. But I'm going to click OK. And when I do that, I get the boxplot. Now, one thing is that by default, it's showing us the ID numbers for the outliers, which is really distracting. So in the next video, I'm going to show you how to modify this to turn off those and make some other changes. But mostly, I want you to see that it's quick and easy to choose boxplot, see what our outcome variable is, which is years at the current address, and then level of education. You can break it down however you want. There are many many other options with the Chart Builder. And when you do one of these charts, you also have options of changing things over here with chart appearance and other options. And there are different ways to modify it as well. I'll show you how all of those work. But this is probably your single best go to for getting started with data visualization and getting an intuitive and an insightful look at what's going on in your data.