This tutorial will introduce you to the Compare panel. The Compare panel contains a series of charts designed to explain the similarities or differences between variables your dataset. To start working with the Compare panel we need a dataset and we need to create a project using it. We will be working with the same Fictional-Employees.csv dataset we used in our Setting Up tutorials. We'll use this to create a project focused on clustering employees based on their characteristic scores.
Once the project is built, we'll dive straight into the Compare panel and start generating some charts. We will use our charts to identify the characteristics that best explain why employees have achieved a High, Medium or Low performance level. Firstly, we will compare the characteristics of high performing employees with all employees in the dataset. Then, we'll move on to compare High, Low and Medium performing employees with each other, saving anything that is notable as an insight along the way.
"Contrast is the intangible ingredient, the catalyst that makes life exciting."
- Maren Elwood
Inside your project's Compare panel, you are asked to choose values from one or more variables. Rather than visualizing the values you have chosen, Graphext uses your choice to create charts visualizing other values in your dataset in relation to the values you have chosen. For instance, comparing female employees with male employees would generate charts visualizing the characteristics of women that most distinguished them from men. In this way, compare charts have been designed to let you quickly recognise the relationships between values in your data or to identify the defining qualities of a value.
If you followed our Setting Up introductions and have already setup a project clustering the Fictional Employees dataset using their characteristics as factors, move onto step 2.
Otherwise, make sure you have the Fictional-Employees.csv file downloaded to your computer. Upload this to the workspace of your Personal team in Graphext and select it to bring up the project setup wizard on the right-hand side of the screen. For a more detailed tutorial on datasets, have a look at our Intro to Datasets tutorial.
Choose Employees as your analysis type and then Clustering. Then, inside of the Clusters and Network Creation menu, set the Performance level variable as your target and set the 10 characteristic variables as your factors.
Finally, inside the Network Visualisation menu, set Name as the variable to identify your nodes by and add Gender to the list of pinned variables. That's it. Continue to name your project something like Fictional Employees: Clustering and execute it. Head back to the Projects panel of your Graphext workspace and hold on for a couple of minutes whilst Graphext builds your project.
For a more detailed tutorial on setting up a project, follow our tutorial Intro to Projects.
Open the Fictional Employees: Clustering project and head straight over to the Compare panel. Since this is the first time we have chosen to compare something in this project, you should see an empty window with an enticing looking search box in the middle. This is where we will pick some values to compare.
It's always good to approach the Compare panel with an aim in mind. Setting a goal helps to define what you want to investigate and should lead you to pick the right kind of values to compare. In this tutorial, we want to identify which characteristics best distinguish High performing employees from Low and Medium performing employees.
This makes for a pretty simple initial decision. Click inside the search field and find Performance Level from the menu list. Selecting it will take you inside of the Compare panel where a series of charts have already been generated for you. Take a moment to browse through these charts.
As you can see, there isn't much comparison going on right now. The charts shown inside of the Compare panel only represent employees with a High Performance Level. You can change that by adding another value into your charts. But before you do this, make sure the Representation dropdown sitting above your charts is set to Relative rather than Absolute. Relative representation displays your values proportionally to the selection you've made. When we add in another value we want to see it displayed relatively.
Now, from the top of your Compare panel, select the plus icon to add another value into your charts. Your charts should now be comparing Low performing employees with High performing employees. First, we want to compare High performers with every employee in the dataset.
Select the dropdown menu above the field representing Low performers with an orange circle. From the menu list, choose Everything as the variable you want to compare. This choice is immediately reflected inside your charts as the values representing Low performing employees have been replaced with values representing every employee in your data!
Compare charts are presented in order of relevance. Since your charts currently represent values that explain the difference between High performing employees and all employees, the variables which best explain this difference will be presented first. This means the first charts in your Compare panel represent the variables where the values belonging to High performing employees are most contrasted with the values for all employees.
Switch your charts so that they represent the variables which best represent the similarity between high performing employees and all employees. You can do this using the dropdown menu at the top left of your Compare panel. Click on Difference and select Similarity from the menu list. As you can see the charts that are presented have changed so that the first ones now represent the variables best explaining the similarity between high performing employees and all employees.
Switch the charts back to represent Difference. Let's move on to inspect the Compare panel in a little more detail. The first line of charts present values for Coachability, Passion and Intelligence respectively. Looking closely at the Coachability chart we can see that the blue line spikes around a Coachability rating of 8. Hover over the blue dot above the Coachability rating of 8. You can see that almost a quarter of our high performing employees are scored at 8 for Coachability. This is an interesting finding and could suggest that a strong Coachability score helps employees to perform well.
It's always good to save potentially interesting discoveries as insights throughout the course of our analysis. Click on the three dots from the top right corner of the Coachability compare chart. Then, click Save as insight from the menu list. This will bring up a window where you can enter a name for the insight. Name it something like 25% of High Performing Employees have a Coachability score of 8, then click Save Insight. This chart will now be saved inside your project's Insights panel.
Have a quick look through the rest of your compare charts and notice that there is a scale presented in the top right corner. This scale represents how relevant the chart is to explaining the Difference between high performing employees and all employees. The more white in the scale - the more relevant the chart is. As you can see, most other charts aren't that relevant so let's change up what we are comparing.
Now, we will adjust the values shown in your Compare panel so that we can compare High performing employees with Low and Medium performing employees. From the top of your Compare panel, delete the value for Everything so that it gets removed from your charts. Now select the plus icon twice to add in two more values to your charts. These should automatically represent Low and Medium performing employees.
That has added a bit more color to your charts! Let's inspect the variables shown to determine whether they have any significance when considering the difference between our High, Medium and Low performing employees.
Find the chart representing Intelligence. There is quite a lot of variation between our three categories here. High performing employees seem to have quite a low Intelligence score, whilst Medium and, to some extent, Low performing employees are more weighted towards the higher end of the Intelligence range. Save this chart as an insight.
In fact, combining this insight with our previous insight about High performing employees that have a high Coachability rating could be even more interesting. If there is a correlation between the finding that High performing employees aren't that intelligent with the finding that they are able to be easily coached, this could well be a meaningful discovery!
It would make sense to save these charts in the same insight because the finding makes more sense when both charts are displayed. You can do this by collectively saving the first 5 charts as an insight.
First, let's hide the chart representing Passion. We've already seen this chart and it wasn't that interesting. Select the three dots next from the top right of the Passion chart and choose Hide from the menu list.
Now that Coachability and Intelligence are the first two charts presented in your Compare panel, it makes sense to create an insight capturing the first 5 charts. Next to the dropdown menu for Absolute vs Relative representation, there is an icon allowing you to do this. Select it and name your insight something like The Most Important Variables when Considering Employee Performance Level and save it. Head over to your Insights panel to take stock of the discoveries you've made so far using the Compare panel.
We know that data isn't always clean and simple.
Have a look through these topics if you can't see what you are looking for.