Pre-Conference Workshops

Python Code Optimisation on NeSI

Wednesday, September 4th, 1:00 – 5:00pm

Register here


Workshop Description
Learn how to optimise Python programs so that they run more efficiently on multi-core systems like NeSI’s high-performance platforms. Designed for NeSI users, this half-day, hands-on training will cover:

-Profiling serial and parallel applications using ARM map.

-Writing a Python C extension, and

-Parallelising a Python C extension using OpenMP

The aim of this training is to increase the intermediate to advanced NeSI user’s confidence in identifying and fixing performance bottlenecks so that they can make the most out of their core hours.


Prerequisites
In order to get the most out of this training, it is strongly recommended that you:

-Have a NeSI account

-Can bring your own device

-Are able to login to a Unix computer

-Are comfortable with typing Unix and git commands (e.g you can clone git repositories, and navigate files/ directories)

-Are able to use an editor (like nano or vim for example) to do things like open, edit, save, and close a file.

-Have some knowledge of Python and C/C++

NeSI is excited to be offering this training and we encourage anyone interested to register. If you are not sure if this workshop is for you or have any further questions, feel free to email training@nesi.org.nz.

—————————————————————————————————————————————————————————
Computational text data analysis using R, quanteda, and tidytext Optimisation on NeSI

Wednesday, September 4th

Register here

Computational text data analysis is an exciting field of research not only for digital humanities, but also for social and health scientists in unraveling and articulating meanings embedded in written text either in native text format or transcribed from audio and video interviews. The free and open-source software R and associated packages tidytext and quanteda make it possible to conduct this complex task based on a corpus of text; however,  many researchers find the task daunting as to how analyse a corpus of text to generate insights applying computational thinking to the process. The goal of this workshop will be to develop computational thinking and enable researchers to use R and packages tidytext and quanteda to conduct computational text analysis.   We will use a specified corpus of text obtained from Project Gutenberg texts, and we will use quanteda and tidytext packages to analyse the corpus of the text. We will use a data carpentry approach and enable practitioners (students, social scientists, and anyone interested in the process) to conduct text data analysis. We will use a hosted Jupyter notebook instance and cover the following steps: (1) reading a corpus of text, (2) tokenise the text corpus, and (3) applying dictionaries to conduct sentiment analysis in the text and (4) identify hidden constructs using topic modelling.  In the workshop, we will use a data carpentry approach using a modular approach to design the lessons, sharing resources over the web, and in the session live coding each step, obtaining frequent feedbacks from the participants using formative assessments in the session.

Computational text data analysis is an exciting field of research not only for digital humanities, but also for social and health scientists in unraveling and articulating meanings embedded in written text either in native text format or transcribed from audio and video interviews. The free and open-source software R and associated packages tidytext and quanteda make it possible to conduct this complex task based on a corpus of text; however,  many researchers find the task daunting as to how analyse a corpus of text to generate insights applying computational thinking to the process. The goal of this workshop will be to develop computational thinking and enable researchers to use R and packages tidytext and quanteda to conduct computational text analysis.   We will use a specified corpus of text obtained from Project Gutenberg texts, and we will use quanteda and tidytext packages to analyse the corpus of the text. We will use a data carpentry approach and enable practitioners (students, social scientists, and anyone interested in the process) to conduct text data analysis. We will use a hosted Jupyter notebook instance and cover the following steps: (1) reading a corpus of text, (2) tokenise the text corpus, and (3) applying dictionaries to conduct sentiment analysis in the text and (4) identify hidden constructs using topic modelling.  In the workshop, we will use a data carpentry approach using a modular approach to design the lessons, sharing resources over the web, and in the session live coding each step, obtaining frequent feedbacks from the participants using formative assessments in the session.

Comments are closed