reticulate: R interface to Python
This talk presents the R package reticulate introduced by RStudio in 2018. It allows to import Python modules, source scripts, convert and manipulate objects and use a Python repl in R.
Tags: Artificial Intelligence, Algorithms, Big Data, Deep Learning & Artificial Intelligence, Data Science, NLP, Machine Learning, Visualisation
Scheduled on wednesday 11:30 in room lounge
Jens Bruno Wittek has been working as a data scientist at AKKA Digital (previously known as GIGATRONIK) in Stuttgart for nearly two years. He conducts projects involving predictive maintenance for car parts and for lasers, text classification and others. Before, he got a Master in statistics at Bielefeld University and then was on loan at Daimler TSS. Since 2009 Jens Bruno Wittek gained expert level in R and now is using both R and Python in business projects involving data management, statistics, machine learning, visualisation and more.
Python and R are the preferred languages for data science. In 2018, RStudio introduced its package reticulate and clearly demonstrates that it favours to join forces. Both languages have strengths and weaknesses. Tools to combine the strengths will enable easier collaboration in projects and more possibilities to succeed. Using Python from R gives R users wider access to functions and makes it easier for Python beginners to just run scripts and being able to collaborate in Python projects. The talk will show the possibilities of reticulate: The main part starts with demonstrating the Python interpreter within R. It will show how to source Python scripts as well as install and import modules. Then it will deal with the most important types of Python objects, how they are represented in R and how to further manipulate them. Thereby, a special focus is on using Python for data science. In addition, it will be presented how Conda environments can be created and used from R. A further application will be the creation of reports with Markdown and LaTeX where R and Python can be used within one document and share objects. A last topic is about showing the possibilities for easier development in RStudio (help regarding Python functions, auto completion).