Question

i would like to plot my data in such a fashion, that bars are grouped up?Something like this:

my code looks like this so far:

data = NP.genfromtxt('newfile',unpack=True,names=True,dtype=None)
for i in sample:
 mask = data['name'==sample]
 ax2.bar(pos-0.5,(data['data']*100,label="samples", color="lightblue")

This created several graphs instead of a combined one, though. How do i convert this into a grafic looking like the one i presented above?

Answer 1

Without having all the code listed, I had to make some assumptions about the data and about what you wanted. Thus, I haven’t tested this code with actual data. I have tried to document my assumptions carefully in the code below. If you have problems with what I’ve posted, perhaps you could post the text file mentioned in the first line of your code.

# Assume data is a record array
data = NP.genfromtxt('newfile',unpack=True,names=True,dtype=None)

# Assume 'sample' is a column in the data
sample = NP.unique(data['name'])
num_items = len(sample)
ind = NP.arange(sample)

# The margin can be increased to add more space between the groups of data
margin = 0.05
width = (1.-2.*margin)/num_items

# This list will make each sample data set a different color
# It must be AT LEAST as long as the number of samples
# If not, the extra data won't be plotted
colorList = ['red', 'blue', 'black']
if len(colorList) < num_items:
    print 'Warning: the number of samples exceeds the length of the color list.'

f = plt.figure()

# Assumed the color was supposed to vary with each data set, so I added a list
for s, color in zip(enumerate(sample), colorList):
    num, i = s
    print "plotting: ", i
    # Assumed you want to plot a separate set of bars for each sample, so I changed the mask to 'name'==i
    mask = data['name'==i]
    # The position of the xdata must be calculated for each of the sample data series
    xdata = ind+margin+(num*width)
    # Assumed you wanted to plot the 'data' column where mask is true and then multiply by 100
    # Also assumed the label and color were supposed to vary with each data set
    plt.bar(xdata, data['data'][mask]*100, width, label=i, color=color)