From the course: The Data Skills Survival Kit

Working with different group sizes

From the course: The Data Skills Survival Kit

Start my 1-month free trial Buy for my team

Working with different group sizes

“

- [Instructor] Sometimes when we study information, we end up having different group sizes, but we need to study these subgroups. Let me give you an example. This data set contains people who have identified if they are an analyst based on a definition. So I'll go do a pivot table really quickly, so that I can look at the different answers. So I'll go here to my Self-Identified Analyst. I'll go and grab the Respondent ID. I'm going to set that to a count. Okay, so in my group scenario, I have 36 people that based on the definition said no, and 63 people who said yes. All right, let's look at this in a visualization. I'll just choose a simple Pivot Chart. I'll do a Clustered Column. Now the size difference is not super dramatic, but it is enough to know that we do have different group sizes, and for the point of comparison, I want to see how people who said no to this question responded to a few of the other questions, while people who said yes, well, I want to look at their responses too. So, the different group sizes presents a little bit of a challenge, but nothing we can't conquer. All right, so let me go back to my different group sizes here. I'll go to Insert, I'll choose a PivotTable, I'll click OK. And I want to take a look at the people who answered yes or no, how often they use Excel. So, I'll throw in the how often they use Excel. I'll scroll down and choose the self-identified analyst. Again, I'll grab my respondent ID, and I'll set it to a count. Okay, so this gives me the count of what I'm trying to accomplish. It tells me that they said they use it a few times a month, but no, they didn't describe themselves as an analyst. Someone who uses it every day, but no, they didn't decide that they were an analyst. Okay, so let's take a look at how this reports when we do a basic chart. So again, I'll just go ahead and do a stacked column here and let me expand it for us to look at it. Okay, so this is showing me, again, look at all these different group sizes, right? Like some people said a few times a month, and yes, they're an analyst, or no, they're not. It's not really giving me what I want though. I really want to see the people who said no, how often they use it, and the people who said yes, how often they use it. So let me change my chart type first, and I'll do this on the design tab. I'm going to go and do 100% Stacked Column, see how that looks? All right, I'll go ahead and click okay here. Again, I sort of have the same problem of 100% of the people who said a few times a month, this percentage is yes, this percentage is no. This is not quite getting me where I want to be. So let me switch row and column. Ah, we're getting closer. Okay, let me expand here so we can take a look. Okay, so of the people who said no, they're using spreadsheets just a few times a week, whereas people who said yes, that they are a data analyst, fewer of them are using it per day. Now look at the every day. So obviously every day usage, for someone who said yes, that makes sense, but there's also a pretty good amount of usage for people who don't declare themselves as a data analyst. About once a week, we can see very few people who actually identify as an analyst said that and never makes sense for someone who said no. But notice how never shows up in the category of yes. When you're dealing with different groups and subcategories and even of different sizes, learning how to shape your data so that you can leverage tools like stacked 100% bars can be very, very helpful to start telling the story behind each one of the answers. Again, as a data analyst, you will definitely be dealing with different group sizes pretty regularly. And the answer and the story is really in how each subgroup answered each question. Be sure to play around with the different chart types and switch those rows and columns and look for your story inside those different sized groups.

Contents