1st semestre, 2016

Date

Speaker

Title (click on it to see Repository)

Feb 18

Ignacio Toledo

An introduction to Pandas

Feb 25

Daniel Moser

True-colour rendering and images manipulation using Python

Mar 3

Paul Eliott

SQL databases

Mar 10

Julien Milli and Bruno Dias

Open discussion on Python

Mar 17

Julien Milli

Fits manipulation

Mar 24

CANCELLED

Happy Easter!

Mar 31

Adele Plunkett

Overview of AstroML from/for a non-expert

Apr 7

Frédéric Vogt

The right way to plot: gridspec

Apr 14

CANCELLED

Break to learn R with Dr. Eric D. Feigelson

Apr 21

Bruno Dias

Digest of open questions in our forum

May 5

Elyar Sedaghati

Bootstrapping with Python and other techniques to get uncertainties

May 19

CANCELLED

ESO workshop in Vitacura

Jun 2

Frédéric Vogt

The right way to plot: colorbars

Jun 16

Daniel Moser

from otherswork import wheel: useful Python packages

Jun 30

CANCELLED

We will resume in August.




An introduction to Pandas
Ignacio Toledo

Abstract
Pandas is a useful tool for manipulation of tables or tabular data in Python. It was introduced in a previous coffee by Ignacio and he will give us here a deeper insight into this package.

Repository
-


True-colour rendering and images manipulation using Python
Daniel Moser

Abstract
In this talk I discuss how to generate human perceived colours from electromagnetic visible spectra, and how to create 'natural contrast' images. The quantitative link between wavelength intensities and physiological perception in human color vision was established by the Commission Internationale de l'Éclairage (CIE) in 1931, which is introduced. To test this relation, I show how to synthesize colors from black-body distributions - and comment why we do not see green stars. The origins of colorblindness and how to make figures that are friendly to colorblind people are also discussed. A key point in the representation of brightness distributions is that the dynamic range of CCD/CMOS sensors and imaging devices (LED/LCD screens) are much smaller than the construction of human vision. Photographic techniques, such as HDR (high-dynamic-range) imaging, can be employed to build much more realistic images. Examples are presented using open-source codes written in Python.

Repository
- Presentation in *.pdf by Daniel
- Example files and codes.


SQL databases
Paul Elliot

Abstract
SQL databases are a very useful way to handle/manipulate large datasets. I will briefly talk about the advantages of using an SQL database and how to interact with the data using python (specifically using the MySQLdb package).

Repository
- Presentation in *.pdf by Paul
- Example files and codes.


Open discussion on Python
Julien Milli and Bruno Dias

Abstract
We propose to have an open discussion on:
- the Python boot camp that Antoine M., Dimitri G, Bruno D. and Julien M. are currently organising for the second part of the year.
- the topics you would like to learn about in the next Python coffees. We will come with some suggestions and want to hear your needs and opinion.
- the future of Python in astronomy, and at ESO in particular
All ideas are welcome !

Repository
-


Fits manipulation
Julien Milli

Abstract
We will review how to open, read, write fits files using the astropy.io.fits package.

Repository
- Example files and codes.


Overview of AstroML from/for a non-expert
Adele Plunkett

Abstract
Straight from the AstroML webpage (www.astroml.org): "AstroML is a Python module for machine learning and data mining," with a goal to implement "common tools and routines used for statistical data analysis in astronomy and astrophysics." It makes clever use of Python routines you may already use, including numpy, scipy, scikit-learn, matplotlib, and others. Entire courses can be formulated around the more advanced statistical topics that AstroML can handle. On the other hand, even a python beginner can find AstroML a user-friendly tool for advanced statistical analyses in astronomy. The goal of this presentation is to (broadly) summarize the AstroML python module and the related resources: including the webpage, github repository and the related text book "Statistics, Data Mining, and Machine Learning in Astronomy" (Ivezic, Connolly, VanderPlas & Gray 2014). While I am neither an expert in Python nor in machine learning/data mining, I find AstroML helpful to know where to turn when faced with a tricky statistical problem, or a nice-looking plot.

Repository
- iPython notebook.
- YouTube course for reference (tip by J. Milli).
- Intro to scikit-learn (tip by A. Mérand).


The right way to plot: gridspec
Frédéric Vogt

Abstract
If you’ve ever made a plot in Python, chances are that you used the matplotlib module. While matplotlib is easy to use, it can be somewhat frustrating when constructing more “advanced” diagrams. In this talk, I will present the greatest - and yet most hidden - gem inside matplotlib: gridspec. This sub-module is essential for creating complex plots with multiple diagrams, extremely useful for dealing with colorbars, and just makes one’s life easier in all situations, no matter the plotting task at hand. Best of all, it will only ever cost you three lines of code, of which the clarity and elegance are of legendary proportions.

Repository
- iPython notebook.


Digest of open questions in our forum
Bruno Dias

Abstract
Our mailing list sc-pycoffee is used to inform about next meeting but also it is used as a forum to ask questions that you could not solve using Google, StackOverflow etc. Currently we have some questions that were not answered and I will show them and publish on our webpage to encourage someone who knows the answer to share it with us. From next week on we will be alternating the Thursday coffee break with the new MCMC, with statistical topics organised by Jorge Lillo-Box and Adele Plunkett. As always, contributions and constructive opinions are encouraged!

Repository
- Presentation in *.pdf


Bootstrapping with Python and other techniques to get uncertainties
Elyar Sedaghati

Abstract
There are a countless number of methods and techniques in interpretation of data for mining the desired information. Be it parametric or non-parametric methods, it is paramount to quote realistic and conservative estimates of precision on the derived parameters of the model being investigated. As in everything else in the universe, Python makes this easy for us. I will, very briefly, discuss a few different ways of obtaining these parameter uncertainties, and highlight the importance of bootstrapping technique application as a first stop in the robust determination of those errors.

Repository
-


The right way to plot: colorbars
Frédéric Vogt

Abstract
Inside matplotlib lies an enchanted World of wonders and magic: the World of Colorbars. Together, let us jump feet first down the rabbit hole and lift the greyish veil of Reality to reveal the limitless nature of plots in Python.

Repository
- iPython notebook.
- Colour blindness simulator


from otherswork import wheel: useful Python packages
Daniel Moser

Abstract
Python is language where simplicity matters. Thus, many features are only accessible through built-in modules, called the standard library. Also, new useful packages are continuously written since Python emerged from the free software principle. In this talk, I will make a (very) short introduction to Python's standard library, and share my experience in finding and installing packages from Python community -- biased for use in astronomy.

Repository
- Code and notes (github)