index.xml

<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Welcome to my blog</title>
    <link>https://liuyanguu.github.io/</link>
    <description>Recent content on Welcome to my blog</description>
    <generator>Hugo -- gohugo.io</generator>
    <language>en-us</language>
    <lastBuildDate>Sun, 15 Dec 2024 00:00:00 +0000</lastBuildDate><atom:link href="https://liuyanguu.github.io/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>2024 National K-12: Does a Higher Rating Lead to a Higher Standing?</title>
      <link>https://liuyanguu.github.io/post/2024/12/15/2024-national-k-12-does-a-higher-rating-lead-to-a-higher-standing/</link>
      <pubDate>Sun, 15 Dec 2024 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2024/12/15/2024-national-k-12-does-a-higher-rating-lead-to-a-higher-standing/</guid>
      <description>Statistics of 2024 National K-12 Grades Championship
https://www.uschess.org/results/2024/k12/
Update: An interactive chart of all players has been added
Over 2,600 contestants from 43 states participated in the 2024 National K-12 Grades Championship
The 2024 National K-12 Grades Championship attracted a total of 2,679 participants from 43 states. More than 1,000 participants (40%) came from New York, followed by Virginia with 250 participants (9%) and New Jersey with 218 participants (8%).</description>
    </item>
    
    <item>
      <title>How many children under 5 worldwide?</title>
      <link>https://liuyanguu.github.io/post/2024/12/13/how-many-children-under-5-worldwide/</link>
      <pubDate>Fri, 13 Dec 2024 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2024/12/13/how-many-children-under-5-worldwide/</guid>
      <description>Since these figures are not directly available via Google search, I calculated them using the mid-year population estimates from the World Population Prospects (WPP) 2024:
World Population: In 2024, the global population is estimated at 8.2 billion (8,162 million). The population is projected to peak in 2084 at 10.3 billion.
Children Under 5 Years Old: The number of children under 5 is 647 million in 2024. This figure peaked in 2017 at 698 million and is projected to decline to 550 million by 2100.</description>
    </item>
    
    <item>
      <title>Plot multiple countries on the world map</title>
      <link>https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/</link>
      <pubDate>Mon, 29 May 2023 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/</guid>
      <description>This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map.
I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries!
The most straightforward way
(You may download the world map from Github)
suppressPackageStartupMessages({
library(&amp;quot;data.table&amp;quot;)
library(&amp;quot;ggplot2&amp;quot;)
library(&amp;quot;rgdal&amp;quot;)
library(&amp;quot;raster&amp;quot;)
library(&amp;quot;rgeos&amp;quot;)
library(&amp;quot;here&amp;quot;)
library(&amp;quot;ggthemes&amp;quot;)
})
# download data from GADM directly and row-bind the spatial polygons data frame
cnames &amp;lt;- c(&amp;quot;Haiti&amp;quot;, &amp;quot;Togo&amp;quot;, &amp;quot;Uganda&amp;quot;, &amp;quot;Ghana&amp;quot;, &amp;quot;South Africa&amp;quot;, &amp;quot;Angola&amp;quot;)
download.</description>
    </item>
    
    <item>
      <title>RMarkdown to Github Pages</title>
      <link>https://liuyanguu.github.io/post/2021/01/06/rmarkdown-to-github-pages/</link>
      <pubDate>Wed, 06 Jan 2021 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2021/01/06/rmarkdown-to-github-pages/</guid>
      <description>RMarkdown into Github Pages
Example of one of my Github pages:
You might have read this GitHub and RStudio tutorial by searching this topic. It is quite long and confusing as it tries to teach Git at the same time. You don’t need any of those branch operations. And the example yaml code is not indented correctly (corrected below).
Here is the short version:
Create in the root directory a yaml file: &#34;</description>
    </item>
    
    <item>
      <title>US and China Admin1 COVID19 mortality and incidence rate</title>
      <link>https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/</link>
      <pubDate>Sun, 14 Jun 2020 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/</guid>
      <description>US
China
Compared to lung cancer mortality
Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable.
Code hosted on my Github repo.
US
# major function, can download from Github repo Blogdown/hugo-xmag/Code
source(here::here(&amp;quot;Code/COVID_make_map.R&amp;quot;))
# US
dt_JUH_US &amp;lt;- get.JHU.us.state()
make_heatmap(data = dt_JUH_US, geo_data = get_state_name(),
state_var = &amp;quot;Province_State&amp;quot;, fill_var = &amp;quot;Mortality_Rate&amp;quot;, label_var = &amp;quot;abb&amp;quot;)
us_maps &amp;lt;- lapply(c(&amp;quot;Mortality_Rate&amp;quot;, &amp;quot;Incident_Rate&amp;quot;, &amp;quot;Testing_Rate&amp;quot;, &amp;quot;Hospitalization_Rate&amp;quot;),
make_heatmap, data = dt_JUH_US, geo_data = get_state_name(),
state_var = &amp;quot;Province_State&amp;quot;, label_var = &amp;quot;abb&amp;quot;)
plot_grid &amp;lt;- gridExtra::grid.</description>
    </item>
    
    <item>
      <title>ggplot US state and China province heatmap</title>
      <link>https://liuyanguu.github.io/post/2020/06/12/ggplot-us-state-and-china-province-heatmap/</link>
      <pubDate>Fri, 12 Jun 2020 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2020/06/12/ggplot-us-state-and-china-province-heatmap/</guid>
      <description>1. US Map by state Method 1. Use usmap Method 2. Use map_data and build from shape files 2. China map by province Method 1. China map by province using downloaded shap files Method 2. Using geojsonMap (leaflet) (Updated on 2020-06-12, First posted on 2019-04-17) It sounds easy but turned out not as straight-forward as I thought. I will show: 50-state (including Alaska and Hawaii) United States thematic map, with</description>
    </item>
    
    <item>
      <title>Working with 3D array as long-format data in R</title>
      <link>https://liuyanguu.github.io/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/</link>
      <pubDate>Sat, 11 Jan 2020 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/</guid>
      <description>Example: a 3-D array with dimension 4x3x2
Melt into long
Recover the array
check.and.install.pkgs &amp;lt;- function(pkgs){
new.packages &amp;lt;- pkgs[!pkgs %in% installed.packages()[,&amp;quot;Package&amp;quot;]]
if(length(new.packages)) install.packages(new.packages, dependencies = TRUE)
suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE)))
}
check.and.install.pkgs(c(&amp;quot;data.table&amp;quot;, &amp;quot;reshape2&amp;quot;, &amp;quot;scatterplot3d&amp;quot;))
Happy New Year!
Recently I spent some time working with array in R.
I believe it is a bad idea to work with array using for loop, which is both slow and error-prone. We can just melt it into a long data, do the work, and arrange back into array in the end if needed.</description>
    </item>
    
    <item>
      <title>Drake: powerful tool for automatic reproducible workflow</title>
      <link>https://liuyanguu.github.io/post/2019/09/15/drake-powerful-tool-for-automatic-reproducible-workflow/</link>
      <pubDate>Sun, 15 Sep 2019 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2019/09/15/drake-powerful-tool-for-automatic-reproducible-workflow/</guid>
      <description>drake is a powerful tool for automatic reproducible workflow. I found it a perfect match when used together with RMarkdown. There are great documentations online for drake thus here I only show a simple example working with RMarkdown.
RMarkdown file could be very slow to generate if lots of calculations are involved. Any small revise makes you rerun everything. When we use drake we can do all the calculations in advance thus the rendering is super fast since we only need to re-do the revised object.</description>
    </item>
    
    <item>
      <title>Notes on writing an R package</title>
      <link>https://liuyanguu.github.io/post/2019/07/28/some-experience-on-writing-r-package/</link>
      <pubDate>Sun, 28 Jul 2019 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2019/07/28/some-experience-on-writing-r-package/</guid>
      <description>Some of my own experience
On description
Namespace
load vs attach
Documentation
Potential problems when checking the package
Some nice suggestions from the CRAN team when submitting the package
Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either
install.packages(&amp;quot;SHAPforxgboost&amp;quot;)
or
devtools::install_github(&amp;quot;liuyanguu/SHAPforxgboost&amp;quot;)
Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package.</description>
    </item>
    
    <item>
      <title>SHAP for XGBoost in R: SHAPforxgboost</title>
      <link>https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/</link>
      <pubDate>Thu, 18 Jul 2019 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/</guid>
      <description>The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance
SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot
References
The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
Please install from CRAN or Github.
install.packages(&amp;quot;SHAPforxgboost&amp;quot;)
# or devtools::install_github(&amp;quot;liuyanguu/SHAPforxgboost&amp;quot;)
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.</description>
    </item>
    
    <item>
      <title>Shiny in Blogdown</title>
      <link>https://liuyanguu.github.io/post/2019/02/24/shiny-in-blogdown/</link>
      <pubDate>Sun, 24 Feb 2019 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2019/02/24/shiny-in-blogdown/</guid>
      <description>How to embed Shiny
My Shiny app example
How to embed Shiny
Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
The only way to do it is using iframe and write outside the chunk:
&amp;lt;iframe src=&#34;https://liuyanguu.shinyapps.io/bcl_app/&#34; width=1000 height=800&#34;&amp;gt;&amp;lt;/iframe&amp;gt;
There is also a built-in function in knitr to do the same thing and write in the chunk. The only difference is that we can only set height, and the shiny app would appear slightly different with the siderbarPanel at the top and the mainPanel beneath it.</description>
    </item>
    
    <item>
      <title>Study shrinkage and DART in xgboost modeling using a simple dataset</title>
      <link>https://liuyanguu.github.io/post/2018/11/15/xgboost-dart-example/</link>
      <pubDate>Thu, 15 Nov 2018 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2018/11/15/xgboost-dart-example/</guid>
      <description>Data
Shrinkage
DART: Dropout - MART
skip_drop
rate_drop
one_drop
It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model.
Data
set.seed(123)
n0 &amp;lt;- 5
X &amp;lt;- data.frame(x1 = runif(n0), x2 = runif(n0))
Y &amp;lt;- c(1, 5, 20, 50, 100)
cbind(X, Y)
## x1 x2 Y
## 1 0.</description>
    </item>
    
    <item>
      <title>SHAP Visualization in R (first post)</title>
      <link>https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/</link>
      <pubDate>Sun, 14 Oct 2018 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/</guid>
      <description>Example 1
SHAP summary plot
Alternative ways:
SHAP dependence plot
SHAP interaction effect plot
SHAP force plot
Example 2
Summary plot
Dependence plot for each feature
Force plot
Stack plot by clustering groups
Update 19/07/21:
Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R</description>
    </item>
    
    <item>
      <title>Spatial data in R: Dividing raster layers into equal-area rings</title>
      <link>https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/</link>
      <pubDate>Fri, 20 Jul 2018 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/</guid>
      <description>Saturation in ten cities with equal-area rings
R Code for one city
Results for the ring saturations
Average saturation in each ring
Methodology
Original Code
This data visualization example include:
* Import .img file as a raster
* Turn raster into a data.frame of points (coordinates) and values
* Dividing the points into 100 equal-area rings
* Calculate Built-up Area/Urban Extent for each ring
* Turn dataframe into raster</description>
    </item>
    
    <item>
      <title>How to Draw Heatmap with Colorful Dendrogram</title>
      <link>https://liuyanguu.github.io/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/</link>
      <pubDate>Mon, 16 Jul 2018 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/</guid>
      <description>Data
Version 1: Color both the branches and labels
Version 2: color only the labels.
Version 3: If there is no color, and we do not reorder the branches
This data visualization example include:
* Hierarchical clustering, dendrogram and heat map based on normalized odds ratios
* The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend
* Heatmap is made by heatmap.2 from gplots using the built dendrogram</description>
    </item>
    
    <item>
      <title>Catalog of my old blog</title>
      <link>https://liuyanguu.github.io/post/2018/07/01/catalogue-of-my-old-blog/</link>
      <pubDate>Sun, 01 Jul 2018 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/post/2018/07/01/catalogue-of-my-old-blog/</guid>
      <description>Introducing my new blog written solely in R Markdown
When I realized it was so convenient to write blog directly using R Markdown, I searched if there is a specific tool for it. And I found Blogdown, an R package developed by Yihui Xie, who also developed R Markdown.
Catalog
As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018.</description>
    </item>
    
    <item>
      <title>About Blogdown and Hugo XMag Theme</title>
      <link>https://liuyanguu.github.io/about-hugo/</link>
      <pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/about-hugo/</guid>
      <description>Yang: This is an original page enclosed by Yihui Xie in the Hugo XMag theme he wrote. I really like it and use it for this blog. This post gives a great introduction to how this blog works.
XMag is the second Hugo theme I have designed. It is based on my first Hugo theme XMin, and similarly, features minimalism but with a magazine style on the homepage inspired by The Signpost on Wikipedia.</description>
    </item>
    
    <item>
      <title>Yang Liu</title>
      <link>https://liuyanguu.github.io/about-me/</link>
      <pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate>
      
      <guid>https://liuyanguu.github.io/about-me/</guid>
      <description>Thank you for visiting my blog.
I am a statistician focused on child mortality at UNICEF. Prior to my current role, I contributed to environmental epidemiology projects at Mount Sinai Hospital and engaged in urban science research at the NYU Marron Institute. I&amp;rsquo;ve been residing in New York since 2014.
In 2018, I launched a blog to document my projects and share insights from my experience with R programming. As my engagement with the blog and my audience grew, I transitioned to a more advanced platform created with R Blogdown, seeking to leverage its enhanced features for a better blogging experience.</description>
    </item>
    
  </channel>
</rss>