From the course: Learning Data Analytics Part 2: Extending and Applying Core Knowledge

Building basic charts visual

From the course: Learning Data Analytics Part 2: Extending and Applying Core Knowledge

Building basic charts visual

“

- [Instructor] When you're a data analyst, you become comfortable in spreadsheets and databases and looking at datasets in their raw format, but showing other people a set of thousands of records really isn't impactful, and even you won't be able to see the big picture with a thousand records. This is where visualization comes into play in the data analyst role. Most analysts typically think about visualization as the outcome of their analysis, and they save building charts and graphs for the end of the process. Research in cognitive science shows us that humans can spot visual patterns pretty effectively, and nothing says there's an issue more than a gigantic column that isn't supposed to be there. In addition to spotting problems early, when we build basic charts and graphs, it helps our decision-makers communicate what they might want to see in the final visuals. Let's look at two visual variations of the PO counts. Let's start with a pie chart. So I'll go ahead and highlight my customer name and my count of the POs, and I'll do an insert. I'll choose just a regular 2D pie. I'm not trying to build presentation graphics; I'm actually just looking at my data. Okay, I'll go ahead and remove that legend. So now this pie chart pretty much shows me that everything is pretty even. I have to look really deep to see any other variation. Okay, so there's one. Let me go ahead and zoom out and let me try to visualize this again in a line chart to see if it's more compelling. We'll just choose, again, a standard 2D line, and right out of the gate, within the second I placed it on the sheet, I immediately see this dip in the PO count. That dip is not quite as obvious inside that pie chart. Now, as an analyst, this dip makes me want to go study this particular customer and look at their numbers. So I start trying to find, oh here it is. Okay. How does their total amount compared to all the others? That's another type of visual. So let me go and highlight my customer name, choosing total amount. And this time, I'll just go insert a basic column chart. Even if I stack them up and try to look at them, I'm still having to look between two visuals to see any type of picture. So let's try combo. All right, so let me highlight my data here and let's do a combo chart. I'll go to insert, I can choose my combo chart. And I want to go ahead and do a custom. All right, first of all, I want to put my count on a line and I want to do a secondary axis. Now, the reason I'm doing that secondary axis is because the PO count numbers are totally different numbers than the actual total amount. So having them separated out on different axis is helpful for visualizing. Now, I want to change my total amount to a bar. Now I can easily see the line in relationship to the column. I'll go ahead and click okay there. Definitely prefer this visual. Size it out a little bit. Okay, let me move it to its own sheet tab. I'm going to call this Test Combo. Click OK. I like this combo chart. I can see the high points of the total count of POs, and I can see the low points of the POs. I can also see how they work with the different order amounts. For example, I can look at the peaks where I see I have a high count of POs, but the total order amount is a little bit lower. And then I've got those dips where I can see the trend of the total count of PO is going down, but those order amounts are high. This is a great way to visualize our total count of PO against the actual total order amount. And this is probably an artifact that I would take to a decision-maker and discuss. It's important to remember that we might pick what we believe is the best display, but we don't always know up front what our stakeholders or decision-makers are going to be using this data for and what visualization might be impactful for them. There's going to always be a little bit of back and forth before you land on the right type of visual. Exploring your data visually as a means of understanding it is not only a time-saver, but it will also help you when someone else asks you about the information you're showing. It'll also help you to start thinking through how you might effectively visualize this later in dashboards or reports. Any artifact that you can discuss with people who will use this report will help them to communicate what they're really looking for, and you'll really accomplish a lot overall just by building some basic charts.

Contents