7 Ways To Use Groups Effectively

Do you need to segment your data into categories? Or aggregate your data at a higher level? Does knowing which category a metric belongs to determine whether you interpret its value as good or bad?

Use groups to add additional dimensions to your analysis. Groups allow you to:

  • Segment your data into categories or types
  • Calculate aggregate metrics for those types
  • Understand the context of individual values

Groups give you that nuanced view of your data missing in standard charts.

Using Groups in Heat Map Explorer

Groups TabTo use groups, switch to the Groups tab in the lower right corner of Heat Map Explorer and select a column to group by from one of the three drop-downs.

On a heat map, data rows are represented as rectangular “cells”. Groups look for repeating values in a column and arrange cells with the same value within a larger group rectangle. See the Getting Started Guide for more info.

When using the default Window style, top-level groups will have a silver border, while second-level groups will have a dark grey border.

Now that you’ve got your data grouped, what do you look for?

1. Look for patterns of colors

Is one group dominated by the color red? Do some groups have consistent colors while others look like a patchwork?

Look for similar colorsSimilar colors within a group indicate the values are close together, pointing to a trend or pattern. If some cells are the same color, but others are not, step back and let your eyes unfocus a bit.

Look for an overall color. If you don’t see one, the values may not be related, or the settings may need to be changed. Explore different set points for the colors by adjusting the thumbs on the color slider, or change to a different metric using the Color drop-down on the Data tab.

In the heat map to the right, the group in the lower-right is mostly red, indicating Jason has been overrunning his budgets. Mary, whose group is the lower left, has a sub-group that is mostly yellow.

It’s important not to force a pattern if one’s not there. Not all data has patterns. Plus, the absence of a pattern might be an important insight.

2. Look for odd colors

Look for odd colorsDo you have a group with all the same color except for one or two cells? Is there a sub-group that has one color while all the other sub-groups in that group have another color?

Odd colors suggest an anomaly or outlier that should be investigated. A single red cell in a group of green cells may be a problem or opportunity. Hover over the cell and check its value, then compare it to the values of the cells around it.

In the project heat map to the right, the red cell in the upper left group merits more investigation, Why is it red when all the other projects in the NYC group are green?

3. Hover over groups

Do you need to see the aggregate values of categories or types?

Hover over the group header to see the aggregate values of all the cells within that group. To see the aggregate values for the entire data set, hover over the Home button on the navigation trail directly above the heat map.

By default, values are summed based on the individual cells within each group. To change the aggregation formula, press the Configure Fields button on the toolbar or use the Data > Fields application menu. Once the dialog opens, the rollup formula can be changed within the Aggregation group on the General tab.

4. Compare the size of groups to the whole map

Want to understand the relative contributions of each group to your entire data set?

Compare the size of groups to the whole mapBy default, the Squarified map type works like a nested pie chart, giving you a parts-of-the-whole view of your data. The size of each group is based on its percentage of its parent group.

For example, in the heat map to the right, the Technology group represents about 70% of our portfolio, with Healthcare and Services each representing about 12%. You can also how the Technology group gets sub-divided, allowing you to see your distributions at multiple levels at once.

Note that you must be using the Sum rollup formula (the default) for the Size metric. Other aggregations are only appropriate for comparing the size of groups to each other, not to the whole data set.

5. Compare the size of groups to each other

Compare the size of groups to each otherAre you looking for the largest or smallest group? Do you need to see how groups are distributed in their size?

The Group Bar map type allows you to compare groups against each other using a nested bar chart. Switch to it using the Group Bar Map icon on the toolbar or using the Format > Map Type application menu. Double-click to zoom into any bar to see a bar chart of the sub-groups within that group.

In the heat map to the right, you can see Local Phone has over twice the units sold as Long Distance.

6. Compare the size of individual cells to each other

Compare the size of individual cells to each otherAre cells mostly the same size or do you have a few large cells and a bunch of small ones?

The distribution of sizes tells you how the values of your size metric differ. If your values are evenly distributed, you’ll see a range of sizes; otherwise larger or smaller cells will dominate.

To more easily compare the sizes of individual cells, switch to the Bar map style. Each cell becomes a bar in a bar chart, with one chart per group. Switch the Label field to the same field you have mapped to Size to have the bars sorted by size, giving you a size distribution curve made up of the individual bars.

The heat map to the right maps Highway MPG to size. Notice how the bars in the Pick-Ups group are mostly the same size, while the bars in the Car group start small, but increase dramatically. You can see here that there’s little variation in highway MPG for pick-ups, but a great deal for cars.

7. Benchmark cells against the group average

How do the values you are mapping to color compare to the average of each group?

Benchmark cells against the group averageTo see, right-click on any cell and select Group Benchmark. The average of the color field is calculated for each group, then the percent difference from this average is calculated for each cell in the group. Cells are colored based on this benchmark, with cells above the average colored green and cells below the average colored red.

If a group has its values all close to the average, it will contain only dark colors. Groups with values far from the average will contain bright reds and greens.

If a group is mostly green or mostly red, as in the heat map to the right, a few items are skewing the average. This is clearest in the upper right group, which has one bright green project while all the rest are red.

Use group benchmark in combination with the size field. Look to see if your large cells are mostly one color, which might indicate a relationship between your size and color fields.

Next Steps

Have you learned how to use groups to improve your data analysis? Let me know if you’ve found this post useful and how you’re using groups in the comments below.

If you liked this post, you may also like: