Tables allow us to easily see how 2 variables (for example "Employee" and "Time") correlate. For me the main benefit of this view is that you can easily find answers to two kinds of questions. Like:
My question: is there a way to create an equally easy to use view that does the same for 3 variables (for example "Employee", "Time" and "Project")?
While it can be computationaly and mathematically complex, the concept of heat maps can help you out here.
The point of a heat map is you add a third dimension to a 2 dimensional plot not by actually having an additional axis, but by using color. For example, your X axis could be "employees", your Y axis could be "project", and time could be represented by color along the XY plane. One color could can mean "more time" while another color can mean "little time".
In the fictious example below, a certain epmloyee has spent a high amount of time on a certain project, less time on closely related projects and no time on unrelated projects.
By using additional colors, a 4th or 5th dimension can be added. For example, all hues of blue can mean "time", whereas all hues of red could mean "resource spent". by combining these colors, you get purple. hues of purple can mean time & resources spent. In my opinion, the complexity (both for a developer as for a user) tend to increase dramatically by each additional dimension added. I don't have an official study to back up that claim, it's just my own experience.
There was a similar question posted a little bit back about visualization combinations (see Best visualization for combinations). In addition to the heat map posted above you also have a couple other choices which are similar but distinct.
A contour plot is very similar to a heat map except that it displays the boundary of the regions. There is a bit of variance in whether or not folks plot the regions with color simply display the boundary. The option gets its name from topographical maps where it truly is showing the contour of the land. It works equally well for displaying 3 pieces of data that are related.
This option is just an extension of the standard 2d scatter plot. Instead of just placing a dot in the region for the data point, you encode additional information in the marker. Things that can be used include:
If you are looking to show relationships between pairs of multiple variables, you can also use the matrix or multiple scatter plot approach. This works well for finding relationships between pairs of large groups of variables.
The parallel coordinates plot works well when you have more than 3 variables to show, but it can work well for 3 also. Its specialty is that it allows for filtering and interaction which can be used to select data for other visualizations. It is also well geared toward encoding color and other data across the lines. It is very compact in space as well considering the number of variables that can be displayed. It is less useful for relationships unless they come in groups.
