PyCon Australia 2013

Analizando las menciones sobre ciencia en Twitter con ayuda de IPython Notebook y Python Pandas

Brenda Moon  · 

Transcripción

Extracto de la transcripción automática del vídeo realizada por YouTube.

hello welcome to afternoon tea onwards our first presenter we have Brenda moon she is a PhD candidate in science communication at the ANU where she is using data mining on social media to understand public attitudes towards science and technology she is a

keen supporter of open source open hardware in open science to talk to us about exploring science on twitter with ipython notebook and python pandas please welcome Brenda moon hello I'm Brenda moon and I'm doing my PhD insights communication at a and

you looking at how people were used the word science on Twitter my initial data collection was done using Java but after trying a few different options for data analysis i settled on pie I Python notebook and Python pandas as my main tools today I'm going

to give you a quick overview of how I've set up my research environment and then show how I play third notebook and pandas help my research I'm fairly new to python so I'd really appreciate any feedback about ways i could be doing things better

for research being able to go back to something you've done a year ago run it and get the same result is really really important but the tools i'm using are improving very quickly and i'd like to be able to take advantage of these improvements

i use python virtual environments to help me achieve these goals virtually inflicts you have a virtual environment on python and probably a lot of you have already heard of it but i've always been surprised at how many people are not using them when they're

so easy to use virtual in forever makes virtual end easier to use and using virtual in forever has made trying out different python modules much easier and it lets me take advantage of the rapid development of the tools that i'm using without worrying

about my coat stopping working I make a note of which virtual environment i'm using at the top of each ipython notebook so that when I go back later on I can know if it stopped working in my new environment that I'm now using I can know which one to

go back to to run it again and then perhaps bring it up to date or if I'm just checking something just run it in the old environment and part of doing that is that copy virtual environment lets you copy your current environment to a new environment before

you install any new software or update any software so that way you can revert to your early environment environment if the latest bleeding-edge version of pirate pandas or at least when I was preparing this talk the latest bleeding-edge version of ipython

wasn't able to start an I Python notebook there was just one little bug and you can revert to your previous environment without you know having to go through a lot of work and even just delete the one that was broken another useful command set virtual

end of project lets you set the virtual lets you set the directory to be used within the particular virtual end so that means then when you activate that virtual environment it automatically changes to the right directory for that project and I find that really

just a useful extra time saving thing as well the next part of my research tool chain I'd like to introduce is ipython notebook I was wondering how many of you have seen ipython notebook and how many of you are using it right so quite a few have seen it

but not a lot of using it yet ok so I'll lips next I'll start up ipython notebook and have a look at the parts of it that I find useful for a research ipython notebook is started at the command line I hope that's visible up there first I activate

the project's virtual environment we're using the work on command and that also navigate to the project directory as I said before the next command shows how I start ipython the pilot parameter tells it to load pylab support and in line makes it show

the matlab metal plot lib graphs on the web page here's what the ipython notebook service had died up looks like in the terminal window I've zoomed in a little bit just so you can see the detail on the projector once it started it opens the notebook

index page in your default browser so i just wanted to show that so you sort of realize that it's not all happening in the browser there is this back end part that you're running that's then starting up the front end part oops wrong button the

first page it loads is an index of the notebooks in the current directory we can create a new notebook drag an existing notebook onto the window or open one that is already in the directory in this case the shutdown button shows that the example notebook is

open and has a connection to the back end notebook server this is what a notebook looks like in this example I'm typing in the main heading so you can see that the current cell type is set to heading 1 a notebook consists of cells and each cell can have

[ ... ]

Nota: se han omitido las otras 2.339 palabras de la transcripción completa para cumplir con las normas de «uso razonable» de YouTube.