many eyes many eyes
Can't delete this data set? That's because it has either visualizations or comments attached to it.
Can't delete this? That's because it has comments attached to it.
Sign in
  • explore
    • visualizations
    • data sets
    • comments
    • topic hubs
  • participate
    • register
    • create visualization
    • upload data set
    • create topic hub
  • learn more
    • quick start
    • visualization types
    • data format & style
    • about Many Eyes
    • FAQ
    • blog
  • contact
    • contact us
    • report a bug
  • legal
    • terms of use

Matrix Chart

When to use a matrix chart

A matrix chart summarizes a multidimensional data set in a grid. For instance, if you had data on contributions to political candidates categorized by candidate and state, you could portray how much money each state gave to each candidate.

Matrix Chart

How a matrix chart works

The matrix chart divides the screen into a grid. Rows represent the values in one text column (e.g., political candidate) and columns represent another text column (e.g., states of the US). Each cell then shows a circle or bar that represents the value for its row/column combination (e.g., contribution to Hillary Clinton from New York).

Bars and bubbles

The two modes of the matrix, circles and bars, are useful in different situations. Bars, whose height represents numeric values, are better for exact comparisons and allow space for more columns. Circles, which show values via area, are good for showing non-negative values that vary greatly, and allow space for more rows. You can flip between the two modes in the "Expert Options" menu to the left of the visualizations.

Expert Options

Colors

You can display a third dimension of data using color. When you select a column to be represented by color, circles will turn into miniature pie charts and bars will be broken into differently-colored pieces. This option should be used with caution, since it's possible to create a cluttered display, but it can often be effective.

Average, sum and count

The matrix chart automatically performs simple aggregation of the data set. For instance, if your data set has two rows that show contributions to Hillary Clinton from New York, say $1,000 and $2,000 the grid cell for Clinton/NY will aggregate these values. If you choose "Sum" in the options pane, the value will be $3,000. Average would yield $1,500, and Count would show 2, the number of rows that correspond to that cell. Note that if your data set has no numeric columns, the only option is to aggregate by count.

Percentages

Sometimes you may wish to compare data within columns or rows. For example, you may not care about the absolute amount of money given by each state, but rather the proportions of each state's money that has gone to each candidate. Use the percentage buttons to compare show proportions in individual columns, rows, or (if you're using color) even within cells.

Interaction:

To see the exact value of a bubble or bar, move your mouse over it. In color mode, you can click on the color rectangles in the legend to customize the colors.

Highlighting:

To highlight an item, click it. Click outside the chart area to turn highlights off. Your highlights will be saved with any comments you make, so you can easily refer to particular parts of the histogram.

Data format:

A matrix chart takes a table with at least two text columns, for the x- and y-categories. If there is a third text column it can be used to determine colors. If there are no numeric columns, then the bubbles or bars simply show the count of its in each category combination.

CandidateContributor NameContributor ProfessionLocationAmount
DumbledorePotterStudentHogwarts1,000,000
DumbledoreGrangerStudentHogwarts1,000
DumbledoreKrumQuidditch PlayerGermany10,000
VoldemortQuirrellTeacherHogwarts5,000

In this table, Candidate, Contributor Name and Contributor Profession, and Location could all be mapped to rows, columns, or colors.

Expert notes

A matrix chart can be seen as the visual equivalent of a "cross-tab" or pivot table.