notebook.community



In [1]:

    
import os



In [2]:

    
import numpy as np



In [3]:

    
%matplotlib inline
import matplotlib.pyplot as plt



In [4]:

    
from astropy.cosmology import Planck15 as cosmo



In [5]:

    
import isotropy

Read in the pickle file



In [6]:

    
mockDataFile = os.path.join(isotropy.example_data_dir, 'snFits.p.gz')



In [7]:

    
sampleData, totalSN = isotropy.read_mockDataPickle(mockDataFile)



In [8]:

    
sampleData.head()



In [9]:

    
# Total number of SN in the simulation (before we threw away bad points)
totalSN









    Out[9]:





1622



In [10]:

    
sampleData['mu_err'] = sampleData.mu_var.apply(np.sqrt)



In [11]:

    
sampleData.head()



In [12]:

    
# mu_err > 5.0 is pretty useless, throw them



In [12]:

    
sampleData = sampleData.query('mu_err < 5.0')

Get statistics of binned quantities



In [13]:

    
df = isotropy.binnedDescStat(sampleData)



In [14]:

    
# removes the nans



In [15]:

    
df = df.dropna()



In [16]:

    
df



In [17]:

    
# Now create samples of size numSN = 50000

Create a sample of sn with z, mu, mu_err



In [18]:

    
numSN = 50000
dfmu = isotropy.drawSamples(df, numSN, totalSN, np.random.RandomState(0))



In [19]:

    
dfmu.head()



In [20]:

    
fig, ax = plt.subplots(1,3)
_ = ax[0].hist(dfmu.z, bins=np.arange(0., 1.4, 0.1), histtype='step')
_ = ax[1].hist(dfmu.mu_err, bins=20, histtype='step')
_ = ax[2].errorbar(dfmu.z, dfmu.mu, yerr=dfmu.mu_err.values, fmt='o')
_ = ax[2].plot(dfmu.z, cosmo.distmod(dfmu.z), 'k-', lw=2.)
ax[0].set_xlabel('x')
ax[0].set_ylabel('Num')
ax[1].set_xlabel('mu_err')
ax[2].set_xlabel('z')
ax[2].set_ylabel('mu')









    Out[20]:





<matplotlib.text.Text at 0x10cfa3750>

These show:

- z histogram in the first panel
- mu_err histogram in the second plot
- hubble diagram in the 3rd plot, with the line showing the true cosmology values

The distribution of mu_err in the 2nd panel is a problem because it is going below 0. Surely not a good repesentation!



In [22]:

    
from utils import plotutils as pl



In [33]:

    
fig, ax0, ax1 = pl.settwopanel(setdifflimits=None)
ax0.errorbar(dfmu.z, dfmu.mu, yerr=dfmu.mu_err.values, fmt='.')
ax0.plot(dfmu.z, cosmo.distmod(dfmu.z), 'k-', lw=2.)
ax0.set_xscale('log')
ax1.errorbar(dfmu.z, dfmu.mu - cosmo.distmod(dfmu.z).value, yerr=dfmu.mu_err.values, fmt='o')
ax1.set_xscale('log')
ax0.set_xlim(0., 1.5)
ax1.set_xlim(0., 1.5)









    Out[33]:





(0.10046954761925471, 1.5)

	z	mu	mu_var
10005	0.968958	14.448511	0.253021
10024	0.861278	14.663764	0.276984
10034	0.820012	14.292957	0.301233
10059	1.040996	15.308183	0.537081
10087	0.981252	14.371889	0.393810

	z	mu	mu_var	mu_err
10005	0.968958	14.448511	0.253021	0.503012
10024	0.861278	14.663764	0.276984	0.526293
10034	0.820012	14.292957	0.301233	0.548847
10059	1.040996	15.308183	0.537081	0.732858
10087	0.981252	14.371889	0.393810	0.627543

	count	mean	std
binindex
1	5	0.639079	0.453628
2	12	0.654472	0.736472
3	53	0.466996	0.407956
4	78	0.382548	0.165145
5	99	0.495683	0.263740
6	145	0.462202	0.166019
7	186	0.589450	0.379160
8	209	0.608655	0.316810
9	221	0.671122	0.260682
10	243	0.762463	0.405684
11	178	0.724115	0.202621

	mu	mu_err	z
0	38.935769	0.355757	0.154881
1	39.435058	-0.157277	0.171519
2	39.512421	0.614198	0.160276
3	41.454780	1.214291	0.154488
4	39.243108	-0.016310	0.142365