Rstudio conference7/14/2023 ![]() Learn how to find more information on functions within R packages.Carry out data manipulation using simple steps to solve complex problems. ![]() Import data into R using csv files (other data sources will be discussed).Take charge of their workflow using RStudio projects.Learn how to set up some of the accessibility features of RStudio.Interact with R using the RStudio environment.Recognise and understand key terms often used by users of R.See examples of R in producing healthcare related visualisations and use publicly available healthcare data to answer questions.You will work with, or have an interest in, data.Ībout the course: This is a great opportunity for anyone who is interested in data analysis using Excel, SQL, Access or uses statistical programmes like SPSS to get started with R and R Studio. For example, you will be able to navigate to drives and files and be confident in searching for solutions to technical problems on the internet. Pre-requisites: No prior knowledge of R is required however, it is assumed that you will have a moderate level of computer literacy. Tidyclust for clustering (unsupervised learning).Learn about how to import data, wrangle data and produce and export graphs. Survival analysis from the survival package. Some additional tools (extensions for tidymodels), from other talks, include: I also liked the method in which tidymodels makes it really easy to tune your models by setting tuning on feature engineering parameters or on the model’s parameters (package tune). The recipe package is pretty rich in data engineering tools, a good reference is the following book on feature engineering and selection: /max/FES. It will be interesting to try and upload a plumber api to an AWS Lambda function using Docker (the vetiver package is able to generate a dockerfile, see here and here for more information). It uses the butcher package to make the model slimmer, removing unnecessary data. Interesting thing to see in the workshop was the vetiver package for MLOps – it makes it easy to generate a model in production (a plumber api). The workshop contained a lot of information and examples and was overwhelming to consume in just two days! Definitely a lot more to try out and learn independently. So far I didn’t do much work in tidymodels, but after the workshop I see that it does seem like a very good unifying framework. I took the opportunity of the conference to join the tidymodels workshop. For more information abot shiny for python, read here. The shinyapps.io service will also support shiny, there will be a shiny server for python, and additional surprises (e.g., shiny for python can be deployed without a server, using static html and by translating the python code to WebAssembly). Joe also told the story of how shiny came to be, how R “fits like a glove” to shiny (due to way function arguments are passed), and eventually surprised the audience with the fact that shiny for python is under way, and already in an Alpha version. Joe Chang announced this in his very interesting (and emotional) keynote. ![]() Quatro is already integrated to RStudio IDE in its more recent versions. ![]() It can generate almost anything that RMarkdown can generate (and more coming in the future). It also has extensions to common IDEs such as VS Code. Quarto support Jupyter notebooks as well as RMarkdown. Those who are familiar will see that it is very similar to RMarkdown, however since it is a standalone software, it does not require R. The development of quarto, a program (command line interface actually), which generates documents. Packages which are developed with R and python in mind, such is the vetiver package which allows deployment of tidymodels. See website here, and scroll down – there is a video of Hadley Wickam explaninng the brand change. The new brand will sound more “natural” to non-useRs (or at least not cause aversion). Re branding – RStudio is changing its name to posit. This is reflected in multiple directions presented during the conference: We are embracing multi-lingual data science… we want to make scientific communication better for everyone. This will allow it to sell its products in a “language agnostic” manner. To my understanding, RStudio has figured out that in order to grow significantly, it must put more emphasis on developing python products directly. For a long time now, RStudio has been developing products for R and Python (e.g., via the reticulate package which allows for using python from R, the support of the IDE for python scripts, and more).
0 Comments
Leave a Reply. |