Interpreting Treemap Visualizations

Interpreting Treemap Visualizations


This learning module supplies the example
used in demonstrations, so that you can follow along with the video instructor. Click the
download button and extract the demonstration example assets from the resulting zip archive
file. In this learning module, we will consider
the ways in which treemaps may be used to visualize data, including their suitability
for evaluating data which has been organized into a hierarchy. As we will see, a series
of nested rectangles represents this data organization and the size and color of each
rectangle can be configured to represent values from the underlying data table. The first thing you should know about treemaps
is that you must have hierarchically organized data in order to make this visualization type
useful. The rectangles in the treemap represent the hierarchically-organized groupings of
data. In this example, four city locations are the highest level of the hierarchy, individual
store numbers are organized within each location, and departments within each store are represented
by colored rectangles. The color and size of rectangles are then set to represent the
magnitude of a continuous variable or variables for the purpose of relative comparisons. In order to appreciate the interactive features
of treemaps, let’s open a configured analysis example in Spotfire. Use Open File and navigate
to the file with Interpreting Treemaps in the title. Click Open, and agree to the terms
and conditions of use. This data table represents values collected
from the sales of products at an office supplies company. Note that sales are organized into
different regions and assigned to one of 76 different sales associates. For each order,
there is information about the price and profit for the products in those orders, as well
as a discount amount which is left to the discretion of each sales associate. We have
over a thousand products which are organized into three primary product categories, seventeen
secondary product categories, and sixty tertiary product categories. And, those columns have
been organized into a custom hierarchy named: Product Hierarchy. It is that information
from the underlying data table which is useful for interpreting the treemap visualizations
which have been created on the pages which follow. On this page, a treemap has been configured
which relates the Average Sale Price as a color gradient, and the sum of Profit values
as a size gradient. There are currently only three rectangles, and both the total profit
and average sales are highest in the Technology product category. Interestingly, though, the
Furniture product category represents the second highest average sales price, according
to the pink color; however, the total profit derived from those sales is the lowest of
the three categories, based upon the small size of the rectangle. As indicated, the true value of the treemap
is its ability to represent hierarchically-organized data. Therefore, let’s slide the hierarchy
slider to the right in order to evaluate the secondary product category level. Notice that
the three primary product categories are still represented by the same-sized rectangles which
allowed us to compare total profit values at that level; however, the color-by variable,
representing the average sale price, has shifted to the lowest level of the current hierarchy
displayed. In addition, the size of these nested rectangles
reflects the total profit for products grouped at the secondary category level. You can see
that while products assigned to the Copiers and Fax category represent our highest average
sale price, the products assigned to Telephones and Communication represent our highest total
profit values. Moving the hierarchy slider to the tertiary
product category level creates a new ‘lowest-level’ color assignment, for comparison of sales
price averages, and nests differentially-sized rectangles within rectangles within rectangles
in order to allow a comparison of total profit values at different levels of the hierarchy.
Sliding to the final level of the hierarchy, where individual products define each rectangle,
the treemap becomes increasingly crowded with nested rectangles. While this may be a valuable ‘high-level’
overview of our entire product line, you may also wish to zoom-in on areas of interest
within the treemap. Each of the ‘parent’ rectangle labels is an active link which,
when clicked upon, will zoom in to that level of the hierarchy. Once you are zoomed in to
the area of interest, labels on rectangles, identifying the current ‘lowest-level’
of the hierarchy, become easier to read. We are also presented with a header across
the top which indicates our zoom path down to this subset of the hierarchy. We can click
on any of these parent levels in order to zoom out to that level, or we can click on
the “(All)” link in order to zoom out completely. So, a hierarchy slider, if presented,
and the ability to zoom in and zoom out are the navigational tools which will allow you
to interpret the information related in a treemap visualization. Let’s proceed to the next tab, where another
treemap has been configured to represent different information. The first thing you may notice
about this treemap is that different categorical variables have been assigned to the hierarchy
property. If we pop up the legend, and widen the dialog in order to read each of the hierarchy
selector labels, you can see the different columns of data which were added, and the
order in which they were chosen to represent this hierarchy. However, because these were added as individual
columns and not a custom hierarchy, there is no slider available to adjust the level
of the hierarchy displayed. Labels for the lowest level of the hierarchy, Order ID, were
turned off, due to the number and size of those rectangles. Note that rectangles change
size based upon the sum of sales, and the color-by variable represents an average of
the discount amount applied to the order. The color property has been configured using
a segmented color mode and a number of points were added in order to define color thresholds
for each segment between the minimum and maximum at values of 2, 5, and 10%. With that understanding of the treemap configuration
provided in this high-level overview of our sales activities, if we want to examine the
details for any areas of the treemap where a preponderance of red rectangles suggests
that the sales associate in that region may have been over-applying discounts, hovering
our mouse over a rectangle will present select values in the tooltip information box.

Related Posts

I Tested The RAREST Paint Colors To Ever Exist…
Kim Kardashian | Draw My Life
How To Draw Earth Day pictures | World Earth Day | Save Earth Images For Drawing Competition

How To Draw Earth Day pictures | World Earth Day | Save Earth Images For Drawing Competition

How To Draw Earth Day pictures | World Earth Day | Save Earth Images For Drawing Competition

2 Replies to “Interpreting Treemap Visualizations”

Leave a Reply

Your email address will not be published. Required fields are marked *