如何使用csv dictreader，matplotlib和numpy创建概率密度函数图？

Question

I'm trying to create a simple probability density function(pdf) graph using data from one column of a csv file using csv dictreader, matplotlib and numpy...

Is there an easy way to use CSV DictReader combined with numpy arrays? Below is code that doesn't work. The error message is TypeError: len() of unsized object, which I'm guessing is related to the fact that my data is not in numpy array format? Also my data has negative and positive numbers. Thanks in advance!

import easygui
import csv
import scipy.stats
from numpy import*
from pylab import*


filename= easygui.fileopenbox(msg='Altitude outlier graph', title='select file',  filetypes=['*.csv'], default='X:\\')
alt_file=open(filename)    

x=[]
for row in csv.DictReader(alt_file):
    x.append(float(row['Dist_90m(nmi)']))

a=scipy.stats.pdf_moments(x)

prob, bins, patches= hist(a, 10,align='left',facecolor='green')

ylabel('probability density function')
show()

Answer 1

The line

a=scipy.stats.pdf_moments(x)

"Return[s] the Gaussian expanded pdf function given the list of central moments (first one is mean)."

That is to say, a is a function, and you must take its value somehow.

So I modified the line:

prob, bins, patches= hist([a(i/100.0) for i in xrange(0,100,1)], 10, align='left', facecolor='green')

And produced this graph with my sample data.

Now my statistics are pretty rusty, and I am not sure if you normally take a pdf over 0-1, but you can figure it out from there.

If you do need to go over a range of floating points, range and xrange do not produce floats, so one easy way around that is to generate large numbers and divide down; hence a(i/100.0) instead of a(i) for i in xrange(0, 1, 0.01) .

Answer 2

Thanks for all the help!! The following code produces a graph of the probability density function: I'm still having some issues formating it but I think this is a good start.

import easygui
import csv
import scipy.stats
import numpy
from pylab import*

filename= easygui.fileopenbox(msg='Altitude outlier graph', title='select file', filetypes=['*.csv'], default='X:\\herring_schools\\')
alt_file=open(filename)    

a=[]
for row in csv.DictReader(alt_file):
    a.append(row['Dist_90m(nmi)'])
y= numpy.array(a, float)    

pdf, bins, patches=hist(y, bins=6, align='left',range=None, normed=True)
ylabel('probability density function')
xlabel('Distance from 90m contour line(nm)')
ylim([0,1])
show()