<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>R on DCoutts Blog</title>
    <link>https://couttsgeodata.netlify.app/tags/r/</link>
    <description>Recent content in R on DCoutts Blog</description>
    <generator>Hugo -- gohugo.io</generator>
    <language>en-us</language>
    <lastBuildDate>Thu, 22 Apr 2021 00:00:00 +0000</lastBuildDate><atom:link href="https://couttsgeodata.netlify.app/tags/r/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Leaky data in geoscience prediction</title>
      <link>https://couttsgeodata.netlify.app/post/2021-04-22-leakydata/</link>
      <pubDate>Thu, 22 Apr 2021 00:00:00 +0000</pubDate>
      
      <guid>https://couttsgeodata.netlify.app/post/2021-04-22-leakydata/</guid>
      <description>Lately, I have been working on a pro-bono geology-focused data science project for a local consulting company here in Calgary. One of the most important things that I learned is how the division of a dataset in to training and test sets impacts model accuracy. Unfortunately I can&amp;rsquo;t share the results of project, so I’ll be demonstrating this concept with anonymized data.
Throughout my learning (Coursera’s data science specialization and other reading) the partitioning of data into training and testing sets was a fairly minor part.</description>
    </item>
    
    <item>
      <title>Let&#39;s do the time warp again!</title>
      <link>https://couttsgeodata.netlify.app/post/2021-03-16-let-s-do-the-time-warp-again/</link>
      <pubDate>Sat, 27 Mar 2021 00:00:00 +0000</pubDate>
      
      <guid>https://couttsgeodata.netlify.app/post/2021-03-16-let-s-do-the-time-warp-again/</guid>
      <description>The Problem - it is difficult to visualize a lot of time series data in a meaningful way Geologist and engineers use time series production data for many tasks. Plotting production volume over time allows them to compare the performance of different reservoirs (or different parts of a single reservoir), as well as look at the structure of the reservoir (e.g., contribution of from primary or secondary porosity). A main problem with time series data is that it has two dimensions that are very costly to visualize as both dimensions are continuous.</description>
    </item>
    
    <item>
      <title>R libraries for geoscience</title>
      <link>https://couttsgeodata.netlify.app/post/2021-02-28-r_geoscience/</link>
      <pubDate>Sun, 28 Feb 2021 00:00:00 +0000</pubDate>
      
      <guid>https://couttsgeodata.netlify.app/post/2021-02-28-r_geoscience/</guid>
      <description>R libraries for geoscience In a chat with another data-oriented geologist, they mentioned a popular R package for geochemistry data (rgr) that I had not heard of. After looking in to rgr a decided to troll through the list of current CRAN packages and find all the geoscience related packages. Again, this is just at list of the libraries that are up-to-date on CRAN, others may have not been kept up to date, but that doesn&amp;rsquo;t mean we can&amp;rsquo;t rally support for them.</description>
    </item>
    
  </channel>
</rss>
