

In [1]:

    
from plotly import __version__
from plotly.offline import download_plotlyjs, init_notebook_mode, iplot

print __version__ # requires version >= 1.9.0



In [4]:

    
init_notebook_mode() # run at the start of every ipython notebook to use plotly.offline
                     # this injects the plotly.js source files into the notebook



In [14]:

    
import plotly.plotly as py
import plotly.graph_objs as go

import numpy as np

t = np.linspace(-1,1.2,2000)
x = (t**3)+(0.3*np.random.randn(2000))
y = (t**6)+(0.3*np.random.randn(2000))

trace1 = go.Scatter(
    x=x, y=y, mode='markers', name='points',
    marker=dict(color='rgb(102,0,0)', size=2, opacity=0.4)
)
trace2 = go.Histogram2dcontour(
    x=x, y=y, name='density', ncontours=20,
    colorscale='Hot', reversescale=True, showscale=False
)
trace3 = go.Histogram(
    x=x, name='x density',
    marker=dict(color='rgb(102,0,0)'),
    yaxis='y2'
)
trace4 = go.Histogram(
    y=y, name='y density', marker=dict(color='rgb(102,0,0)'),
    xaxis='x2'
)
data = [trace1, trace2, trace3, trace4]

layout = go.Layout(
    showlegend=False,
    autosize=False,
    width=600,
    height=550,
    xaxis=dict(
        domain=[0, 0.85],
        showgrid=False,
        zeroline=False
    ),
    yaxis=dict(
        domain=[0, 0.85],
        showgrid=False,
        zeroline=False
    ),
    margin=dict(
        t=50
    ),
    hovermode='closest',
    bargap=0,
    xaxis2=dict(
        domain=[0.85, 1],
        showgrid=False,
        zeroline=False
    ),
    yaxis2=dict(
        domain=[0.85, 1],
        showgrid=False,
        zeroline=False
    )
)

iplot(go.Figure(data=data, layout=layout))
#ipy.plot(fig, filename='2dhistogram-2d-density-plot-subplots')



In [2]:

    
from IPython.display import Image

Dark Energy and Data Science

The nature of dark energy is one of the most pressing questions in cosmology today. A compelling avenue for understanding dark energy is by measuring light coming from high redshift Type 1a (thermonuclear) supernovae. Type 1a supernovae allow us to calculate cosmic expansion because there exists a correlation between the characteristics of the light coming from these supernova, and their intrinsic brightness. Thus, their physical distance can determined in a manner independent of redshift. The first such surveys were critical in the initial discovery of accelerated cosmic expansion. However these surveys were not sensitive enough to definitively select between current cosmological models. Next-generation telescopes, such at the Large Synoptic Survey Telescope (LSST), will not only be more sensitive, but also produce more data, which provides both new challenges and opportunities. This work is focused on understanding the systematics of the statistical modeling techniques that will be used to analyze this wealth of data.

To accomplish this, we work with simulated supernova data in the form of light curves - essentially time vs. luminosity plots. The light curves are passed through an analysis pipeline which fits the data to a parameterized model. This parameterized model was developed and then trained with previous supernova data. As there is some intrinsic scatter in the characteristics of these supernovae, and so few supernova to be used in a training set, there is a bias associated with using this empirical model. Historically, the error associated with model bias was much smaller than the dominant source of noise - light hitting the detector that is not associated with the supernova, called sky noise. However, sky noise follows a Poisson distribution, and so the associated signal-to-noise ratio (SNR) goes as $ 1/ \sqrt{N} $. This means that as data sets get large (as N increases), and the error from sky noise decreases, the errors introduced by model bias become increasingly important.

One particular aspect of potential model bias that we are interested in understanding is how well the model describes the color of the supernova. The color is a measure of how bright the supernova is in different wavelengths, as viewed by the observer. The color is the result of a variety of physical processes, mainly reddening (from dust) on the light's journey through the supernova's galaxy and then through the Milky Way, as well as the intrinic color of the supernova. In the empirical model being used, this is encapuslated with one parameter. We would like to determine how well this model can recover the intrinisic color when host galaxy reddening is not negligible.

Once the analysis code is a complete, this type of analysis will be run on a large data set of simulated supernovae, to gauge the performance of the statistical fits.

Tracing Cosmic Expansion through Type 1a SNe



In [15]:

    
Image(filename='/Users/lisaleemcb/Desktop/muBetoule.png')









    Out[15]:

A Typical Type 1a SN Light Curve



In [3]:

    
Image(filename='sn519LC.png')









    Out[3]:

Fitting Light Curves to Models



In [4]:

    
Image(filename='sn519.png')









    Out[4]:

From Models to Distance Modulus

Next Steps



In [ ]:

Title: Think of this as an advertisement. Very often, people just look at the title to decide whether the poster/talk is interesting. So, let us think about the audience (and the fact that you only have a few words for the title) and ask

Who do we want this to appeal to ?: Cosmology/astronomy/astrophysics folks?: What specifically would appeal to these people Folks from other areas (this would be a large fraction, good majority): What would appeal to these people
How descriptive does this need to be?

Overview:

1st paragraph: I think this is a good level to aim for, if you put in a bit more work you can fine tune it a little better, and squeeze something

I think there are a couple of ideas missing in the overview which would be good to put in:

What is the connection between dark energy/cosmology and light curve fits that you are doing ?
What is the adding host extinction to the list of parameters? Suppose we never investigated this, what would be the potential problems in cosmology?