#A 3.5: ISOMETRIC BAR CHART

Isometric bar charts use the perspective projection of visual elements to display several sets of quantitative data within the same diagram that would be overlapping in a conventional two- dimensional chart. In fact, they represent an alternative style of multiset bar charts instead of breaking the different sets apart and placing groups of bars of the same variable next to each other, they allow the designer to keep the original structure of each subchart.

Although the creation of a perspective structure is more complex than drawing a “fat” bar chart, the isometric view offers several advantages such as a small demand of display space (note that bars can partially overlap each other). However, the effect of complex graphics to attract the user’s attention always bears the danger of producing “chartjunk” instead of effective visualizations (Behrens, 2008).

Use an isometric bar chart to display multiple datasets that consist, similarly to the Multiset Bar Chart pattern, of quantitative data items. Note that you can use several independent datasets, but that you have to make sure that each set contains the same nominal class of discrete items (Behrens, 2008).

Create a three-dimensional Cartesian coordinate system in isometric or axonometric view. Assign the variables of your datasets to the axes. Usually the x- or the y-axis (which both represent the horizontal axes in the three-dimensional chart) holds the identifier of the different datasets you want to display. Make sure shading and alignment are correctly applied to support the perspective view. The two remaining axes act as they would in a two-dimensional coordinate system display ing only one bar chart. The isometric perspective suggests that the different bar charts are placed behind each other (Behrens, 2008).

Isometric projections are easily interpreted by human perception as three-dimensional objects as long as perspective and shading are not completely out of place. Hence, with relatively little effort they allow the designer to add a dimension to its graphics and therefore the possibility of displaying one more variable at a time.

On the other hand the temptation of using perspective visualizations as “eye candy” is quite high, and in many cases an isometric chart looks very attractive indeed. However, the look of an infographic should never be the main reason for opting in favor of such a display technique (Behrens, 2008).

• A 3.1 Simple Bar Chart
• A 3.2 Multiset Bar Chart
• A 3.3 Dot Matrix
• A 3.4 Stacked Bar Chart

from datos import data
d=data('mtcars')
t1 = d.pivot_table(values = 'carb',index=['cyl'], columns = ['gear'],
aggfunc=len)

• Matplotlib
• Numpy

from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
import numpy as np
from datos import data

d=data('mtcars')
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d',  xlabel='Cylindres' ,
ylabel='Gears', zlabel='Frequency', title='Car Distribution by Gear
and Cylindres')
t1 = d.pivot_table( values = 'carb',index=['cyl'], columns = ['gear'],
aggfunc = len)
x,y =t1.index, t1.columns
xpos, ypos = np.meshgrid(x , y )
elements = len(x) * len(y)
xpos = xpos.flatten()
ypos = ypos.flatten()
zpos = np.zeros(elements)
dx= np.ones_like(zpos)
dy=0.4 *np.ones_like(zpos)
dz=t1.values.flatten()
dz[np.isnan(dz)] = 0
c=['salmon','aqua','lightpink',
'salmon','aqua','lightpink','salmon','aqua','lightpink']
ax.bar3d(xpos, ypos, zpos, dx, dy, dz, color=c, zsort='average')
plt.show()

\

counts <- table(mtcars$cyl, mtcars$gear)
counts

##    
##      3  4  5
##   4  1  8  2
##   6  2  4  1
##   8 12  0  2

• lattice
• latticeExtra

d <- table(mtcars$gear, mtcars$cyl)
data<-as.data.frame(d)
names(data)<-c("Gears","Cylindres","Frequency")
data$Gears

## [1] 3 4 5 3 4 5 3 4 5
## Levels: 3 4 5

library(latticeExtra)

## Loading required package: lattice

## Loading required package: RColorBrewer

cloud(Frequency~Gears+Cylindres, data, panel.3d.cloud=panel.3dbars, col.facet='grey', 
      xbase=0.4, ybase=0.4, scales=list(arrows=FALSE, col=1), main="Car Distribution by Gears and Cyl", 
      par.settings = list(axis.line = list(col = "transparent")))