R

So I started seeing things on twitter about using docker to run a standard version of R and with most the libraries I use on a regular basis already installed. I gave it a shot and now I need to sort this out even more – what a great tool! In addition, if you have a group of people using R on a project it should be quite simple to state the image to use or actually provide the image of R, Rstudio, and all the packages in one place.

I have been using [git]() for version control of my various projects including making local backups and versioning as well as for collaborative works utilizing github. Below I provide some simple notes on using git. Basic git To initialize a git project use git init or git clone _some url_ git status shows what has changed since the last update, including if there are differences with a collaborative project online (e.

In a previous post I used SuperLearner for Predictive Mapping. Here following guidance from Hefin Rhys’ youtube tutorial and some scripting made available by Tomislav Hengl of OpenGeoHub. I will demonstrate generating a ranger, random forest, model to predict multiclass and probability of each class. Data is currently closed but I hope to make it open down the road. The mlr package (machine learning in r), is a package that acts as a wrapper for numerous machine learning algorithms.

This is a place where I will place some of the R tricks and script snippets that I need on a semi regular basis… but that I keep forgetting how to do. Filtering character vectors A common example is a need to load a list of rasters from a list of files. The challenge is that the file list can often come with unwanted items in the list (e.g. xml meta-data files).

Why have I never seen this website before?? https://ropensci.org/ It has a great list of packages and a peer review system. Maybe I should join … sometime. I will need to look through their package catelog some more. Here are a few packages that relate to my interests and that I expect to find useful1: Package Author Description bib2df Philipp Ottolinger Parse a BibTeX File to a Data Frame (works both ways) getlandsat Scott Chamberlain Get Landsat 8 Data from Amazon Public Data Sets gitignore Philippe Massicotte Create Useful .

I am really enjoying the power of bookdown to produce great documents that can be read as websites (i.e. html or gitbook output), pdf (generated via LaTeX), or other. However, when you start wanting to format your tables conflicts can occur causing all of some of the book to not render. Below I provide a quick solution. The kableExtra package provides well documented and excellent features for formatting tables into whatever format you are generating.

Upgrading R The following assumes that RStudio is installed. First, of course, download the latest R version from CRAN install.packages("devtools") Install the packages that you need. Here I suggest that you install packages that have lots of dependencies first – as this should install of all of the dependencies. … kind of top down install process. My packages Stage 1: My essential everyday packages devtools bookdown tidyverse bookdown GIS and remote sensing sf raster tmap mapview lidR (see below) rgdal gdalUtil Statistics car lsr nlme – Package to perform linear mixed effects model analysis agricolae – Package for performing Tukey HSD lsmeans – Package for doing pairwise tests wtih lme results moments psych Machine Learning caret SuperLearner mlr3 mlr3learners mlr3viz Model metrics vip vimp Pretty latex2exp rmdformats – alternate reports look for rmarkdown gridExtra Parallel Processing snow doParallel Install CRAN packages plist <- c("devtools", "bookdown", "blogdown", "tidyverse") plist <- append(plist, c("sf", "raster", "tmap", "mapview")) plist <- append(plist, c("car", "lsr", "nlme", "moments", "psych", "caret", "SuperLearner", "mlr3", "mlr3learners", "mlr3viz")) plist <- append(plist, c("vip", "vimp")) plist <- append(plist, c("latex2exp", "snow")) lapply(plist, install.

Authoring Citation Software Zotero: Your personal research assistant. Entering grad studies and beginning many weeks of reading this has revolutionized my life. Better Bibtex for Zotero keep an always up to date .bib file for integration with LaTeX or Rmarkdown citation. LaTeX I’m not totally convinced that I want to commit to LaTeX as Rmarkdown and bookdown do such nice jobs of it. That being said it is a great fallback.

I am currently working on a project that involves machine learning to predict of ecological attributes across landscapes at fine spatial resolution. Our current list is consists of 80 covariates stacked together. To make predictions across the landscape significant tiling of the raster data is needed. However, the tiling process itself takes significant time and creates a tiled dataset in the thousands of tiles… There is another way!

Recently I had the opportunity to attend a machine learning workshop hosted by the provincial BEC Team in Victoria, BC. and led by Tom Hengl of the OpenGeoHub Foudation. During this workshop we worked through various R machine learning packages. I learned some great new skills. The challenge really came down to taking time to understand all the steps – and then to apply it to my data / my projects.