distribution for existing data

Question

I have some data (TEMP_2) and I wanted to obtain a distribution for that data. I know how to do the histogram using:

import numpy as np
from pylab import *

plt.figure(1)
data1 = loadtxt("TEMP_2")

a= data1[:,1]
plt.hist(a,100, normed=True,)
show ()

But, I wanted to have a distribution. Can anybody please help me with this..

data file:

1000 299.23
2000 310.56
3000 308.21
4000 305.86
5000 305.21
6000 301.35
7000 295.37
8000 307.80
9000 295.61
:      :
:      :
200000 307.18

You'll have to be more specific about what you want. The raw data is itself a (very specific) distribution — Adam Rosenfield, Oct 18 '12 at 15:43
I wanted to obtain a distribution plot with out showing the bins. Just , a one line distribution plot — Shila D, Oct 18 '12 at 15:47
As I mentioned before, I don't want to present my data with the histogram.. I mean , the bars of the histogram. — Shila D, Oct 18 '12 at 15:49
First you said you want to have a distribution and plot it, but then you say you don't want to present it with the histogram. I'm confused :/ What do you want? — jsalonen, Oct 18 '12 at 15:51
I need something like this:http://www.google.com/imgres?um=1&hl=en&sa=N&biw=1301&bih=609&tbm=isch&tbnid=80rm0nRHEBZwyM:&imgrefurl=http://glowingpython.blogspot.com/2012/07/distribution-fitting-with-scipy.html&docid=nWXdc4vI56MtfM&imgurl=http://4.bp.blogspot.com/-LKBzu8-e3HA/UAUPkuhGUaI/AAAAAAAAAak/pXHiEoNTfm4/s1600/normaldist.png&w=812&h=612&ei=ZSWAUMCPJOqW0QGKioGACg&zoom=1&iact=rc&dur=552&sig=114445739387104976107&page=1&tbnh=142&tbnw=175&start=0&ndsp=17&ved=1t:429,r:12,s:0,i:108&tx=111&ty=72 — Shila D, Oct 18 '12 at 15:51
You have to tell it what distribution you want to plot. Do you want a normal? What's the mean and stdev? — Joel Cornett, Oct 18 '12 at 16:08

score 1 · Answer 1 · edited May 23 '17 at 10:24

1

Try with:

import matplotlib.pyplot as plt
import numpy as np
import matplotlib.mlab as mlab
from scipy.stats import norm

mean, sigma = norm.fit(data) #your data here
x = np.linspace(-3,3,100)
plt.plot(x,mlab.normpdf(x,mean,sigma))

plt.show()

As described here: python pylab plot normal distribution

edited May 23 '17 at 10:24

Community

1
1

answered Oct 18 '12 at 16:16

luke14free

2,529
1
17
25

Thanks luke14free! .. It works great! – Shila D Oct 18 '12 at 16:33

jsalonen · Accepted Answer · 2012-10-18T16:25:32.237

0

In order to plot a normal distribution that fits your data you need to do the following:

First you need to calculate, which normal distribution best fits your data. In scipy there is norm.fit. Next you just need to plot a normal distribution with the given properties (mean, stdev).

Full script:

# Load data
import numpy as np
from pylab import *
data1 = loadtxt("TEMP_2")
a = data1[:,1]

# Fit data into normal distribution
from scipy.stats import norm
mean, stdev = norm.fit(a)

# Plot normal distribution
import matplotlib.mlab as mlab
x = np.linspace(min(a), max(a), 100)
plot(x, mlab.normpdf(x, mean, stdev))
show()

Result:

Plot screenshot

If you want to plot the "bins" too, then just add before plot.show():

plt.hist(a, len(data1), normed=True,)

edited Oct 18 '12 at 16:25

answered Oct 18 '12 at 16:13

jsalonen

29,593
15
91
109

You're welcome. Please be aware that as I found your question to be very vague, I'm really unsure if this is the correct kind of a distribution plot. – jsalonen Oct 18 '12 at 16:38
Gaussian distribution was the one I looking for..Sorry not mentioning it before... I think we are in the right page... Thanks! – Shila D Oct 18 '12 at 16:52

distribution for existing data

2 Answers2