A new RStudio addin to facilitate inserting tables in rmarkdown documents

In the last months, I started increasingly using Rmd documents for preparing scientific reports, blog posts, etcetera. While I really like the flexibility offered by the system, one thing that I thought could be improved is the support for easily inserting tables. So, “inspired” also by the recent addition of the excellent insert image addin in blogdown, I decided to give it a go and try to implement some kind of addin to facilitate table insertion in Rmd documents.

After struggling a bit due to my rather nonexistent shiny skills, in the end I managed to obtain a “basic but useful” (IMO) addin.

See full post on my new blogdown page here: https://lbusett.netlify.com/post/a-new-rstudio-addin-to-facilitate-inserting-tables-in-rmarkdown-documents/

Automatically importing publications from bibtex to a hugo-academic blog

The problem

I am in the process of migrating my (rather ugly) small blog from “Bloggers” to `blogdown` and, as several others, I choose to use the hugo-academic theme due to its good looks, simplicity, and “focus” towards researchers.

One nice feature of hugo-academic is that it includes out-of-the-box a “Publications” section, allowing researchers to easily create a list of their publication as a section of the website.

Unfortunately, in order to populate that list, users have to manually create one different .md file for each publication, by cutting and pasting several different info (e.g., title, authors, etc.) in a "simple" template like this one.

THIS IS BORING!

Speeding up spatial analyses by integrating `sf` and `data.table`: a test case

The problem

Last week, I replied to this interesting question posted by @Tim_K over stackoverflow. He was seeking efficient solutions to identify all points falling within a maximum distance of xx meters with respect to each single point in a spatial points dataset.

If you have a look at the thread, you will see that a simple solution based on creating a “buffered” polygon dataset beforehand and then intersecting it with the original points is quite fast for “reasonably sized” datasets, thanks to sf spatial indexing capabilities which reduce the number of the required comparisons to be done (See http://r-spatial.org/r/2017/06/22/spatial-index.html). In practice, something like this:

MODIStsp 1.3.3 is out - Speeding things up and squashing some bugs !

  A new version of MODIStsp (1.3.3) is on CRAN as of today ! Below, you can find a short description of the main improvements.

Processing speed improvements

  Processing of MODIS layers after download (i.e., scale and offset calibration, computation of Spectral Indexes and Quality Indicators) is now much faster.


  As you can see in the figure, processing time was almost halved on my (not so fast) laptop. This was achieved by modifying all computation functions so to use `raster::calc()` and `raster::overlay()` (more on this in a later post).

Although speed is also limited by download speed and compression options, this will allow to save quite some time when working on large areas and with many MODIS layers.

Building a website with pkgdown: a short guide

As promised in my last post, here is a short guide with some tips and tricks for building a documentation website for an R package using pkgdown.

In the end, this guide ended up way longer than I was expecting, but I hope you'll find it useful, although it often replicates information already available in pkgdown documentation !

Prerequisites

To build a website using pkgdown, all you need to have is an R package hosted on Git Hub, with a file structure "tweaked" with some functionality provided by devtools. 

Assuming you are using RStudio, and that you didn't already do this, open the project corresponding to your package and (as a minimum) run:

require(devtools)
use_readme_rmd()
use_news_md()
use_vignette("test")  #substitute with the name of your package

The new MODIStsp website (based on pkgdown) is online !

The MODIStsp website, which lay abandoned since several months on github pages, recently underwent a major overhaul thanks to pkgdown. The new site is now available at http://lbusett.github.io/MODIStsp/

MODIStsp (v 1.3.2) is on CRAN !

We are glad to report that MODIStsp is now also available on CRAN ! From now on, you can therefore install it by simply using:

install.packages("MODIStsp")

In v 1.3.2 we also added the functionality to automatically apply scale and offset coefficients on MODIS original values according with the specifications of single MODIS products. Setting the new "Scale output values" option to "Yes", scale factors and offsets are applied (if existing).

In this case, for example, Land Surface Temperature values in the output rasters will be in °K, and spectral indices will be floating point values (e.g., NDVI will be between -1 and 1 instead than between -10000 and 10000).

We also corrected a few bugs, affecting in particular ftp download, and modified the names of some output layers to reduce the length and homogenize output file names, and correct a few  errors.

The changelog for v1.3.2 can be found HERE

We hope you will find the new version useful and that we didn't introduce too many bugs ! Please report any problems in our issues GitHub page.

The `development` version of MODIStsp, containing the latest updates and bug fixes,  will still be available on GitHub. It can be installed using:

library(devtools)
install_github("lbusett/MODIStsp", ref = "master")

"MODIStsp" is a R package allowing automatic download and preprocessing of MODIS Land Products time series - you can find additional information here