Physics
4 (Freshman Seminar) - Fall 2014
Astrophysics and Cosmology with Open Data
Astrophysics and cosmology are in the midst of a golden age of
science-rich observations from incredibly powerful telescopes of
various kinds. The data from these instruments are often freely
available on the web. Anyone can do things like study x-rays from
pulsars in our galaxy or gamma rays from distant galaxies using data
from Swift and Fermi; discover planets eclipsing nearby stars using
data from Kepler; measure the expansion of the universe using
supernovae data; study the cosmic microwave background with data from
Planck; find gravitational waves from binary black hole mergers using
data from LIGO; and study the clustering of galaxies using Hubble
data. We will explore some of these data sets and the science than can
be extracted from them. A primary goal of this class is to develop
skills in scientific computing and visualization - bring your laptop!
Hubble diagrams: then
(Hubble 1929)
and now
(SCP)
The gamma-ray sky as seen by
Fermi
The anisotropy of the microwave sky as seen by
PLANCK
Gravitational waves from
merging neutron stars (left), and
merging black holes (right)
Left:
are spinning neutron stars.
Right: Catalog of pulsars from
Australia Telescope National Facility (ATNF)
Galaxy catalog from the
Cosmic Evolution Survey (COSMOS) project
Data from the Exoplanet Archive.
It goes without saying
that this page is under construction.....
.
Announcements and links:
Check this space often for announcements, rescheduling, etc.
- Instructor:
Prof. Alan J. Weinstein
-
Office: 354A West Bridge, M/C 100-36, x2166
E-mail: ajw AT caltech.edu
I do not have regular office hours; email me!
- Meetings:
Wednesday and Friday (TBD!) at 3pm, in 351 West Bridge
- Feedback:
- I greatly appreciate student feedback; feedback prior to the
end-of-term evaluations lets me modify the class to fit your needs.
In person, by email to
ajw AT caltech.edu,
by campus mail, whatever you like.
If you would like to preserve your anonymity, campus mail will usually work.
My mailbox is on the 3rd floor of West Bridge.
- Outline:
Note that this is way more than what we can get through in one term!
- Set up your Python analysis environment (see below).
- The accelerating expansion of the universe from Type 1a supernovae,
and open data from the Supernova Cosmology Probe project
and the SuperNova Legacy Survey.
- The
cosmic microwave background (CMB), using data from from
PLANCK
and from BICEP-2:
- Worthy links:
- Analysis:
- Download the CMB map from Planck data - it's 2GB!
- It's a FITS file, you will need astropy to read it; it is in the Canopy Package Manager.
- The map uses healpix to pixilate the sky. To read it in python, you will need healpy (python interface to healpix). Unfortunately, not only is this not in the Canopy Package Manager, it will not work on a Windows computer; if that's what you have, you will need run something like linux on VirtualBox.
- On a Mac, in Canopy, Tools -> Canopy Terminal -> "easy_install healpy"
- Healpy has tools
(specifically, anafast)
to analyze the CMB anisotropy map and compute a multipole power spectrum, for comparison to theory.
- Example code to do this: cmbmap.py.
- Analyzing the CMB anisotropy multipole power spectrum
- Legacy Archive for Microwave Background
Data Analysis (LAMBDA),
part of the
High Energy Astrophysics
Science Archive Center (HEASARC).
- Code for Anisotropies in the Microwave Background (CAMB).
Please run the CAMB Web Interface
(or, if you're ambitious, try to download, build, and run the code on your computer;
you'll need a mac, linux, or a linux VM).
Vary the values (choose "Use Physical Parameters?"=No) of Omega_baryon,
Omega_CDM, Omega_Lambda, and Omega_neutrino,
remembering that Omega_curvature is 1-(sum of all the others).
try also varying the number of massless neutrinos from its nominal value of 3.04,
the Scalar Spectral Index from its nominal value of 0.96,
and maybe even the Dark energy Eqn. of State from its nominal value of -1.
For each case (try to do 5 or more), hit "Go!" at the bottom,
then download "camb_xxxxxx_scalcls.dat". Keep a careful record!
Then read in those data and plot them, all on one plot, with labels
to distinguish the cases.
You will see that the "acoustic oscillations peaks" the the temeprature
power spectrum will vary. Which peaks are most sensitive to which
cosmological parameters?
[NOTE that the physics behind the calculations being done here is rather complex;
it won't be easy to understand it with freshman physics!]
- OK, here's a much easier, and maybe more fun, way to do the same thing:
CMB Analyzer.
- A nice
review of the CMB from the Ph10 reading list.
- Gravitational waves (GWs) and their astrophysical sources,
and open data from LIGO.
- More stuff that we will not have time to get to this term:
- High energy gamma ray sources, from the
Fermi space telescope.
- The pulsar catalog, in the radio from the
Australia Telescope National Facility
(ATNF) catalog, and in the x-ray from the
Fermi catalog.
- A galaxy catalog from the
Cosmic Evolution Survey (COSMOS) project.
- An exoplanet catalog from the
Exoplanet Archive.
Python
Set up your computer for python (a nice, powerful interactive language,
that they use in CS1 and that is heavily and increasingly used in scientific computing).
- Install
python,
numpy,
scipy,
matplotlib,
astropy,
h5py.
-
Windows, Linux, Mac:
Enthought
offers some pre-compiled versions of Python that will run with a simple point and click.
A free
academic license
is available to students, staff, and faculty
at academic institutions, and this may provide an easy-to-install solution.
Use your .edu e-mail address to request a free academic license;
Download and install Canopy Express;
Enter your license information to get the full version.
Use the package manager to get all the above packages.
- Browse the
python tutorial,
and the tutorials/FAQ/fast start for each of the other packages.
- If you have a mac, you can install python with
MacPorts
following
these instructions.
You can probably install the latest version (python34);
anything from v2.7 and up is ok, but make sure you get the proper version
of numpy, scipy, matplotlib, astropy, h5py.
-
Linux: use whatever package system is native to your flavor of linux.
- Python for Data Analysis
Last Updated: Nov 26, 2014
Alan Weinstein/ajw AT caltech.edu