The skills I developed learning R compounded. Learning to code improved my computer skills including skills with spreadsheets and swapping out to a Linux box rather than Windows.

When I read this article, I thought of my journey and I thought of guitar players. About a month ago, I had an itch to buy a guitar. My previous attempts were failures. My uncle told me years ago in essence that if you need lessons you don’t have it or at least that is what I took away. I have had a sense that you were born a musician or not. Directionally this may be correct but what we are missing and that what the article is saying is we are missing the effort put forth. That is the born musician still practices and works at their craft.

I researched and became convinced that one could acquire the skill of learning a guitar. My goals were to put my focus on something other than work and to get off the computer/phone.

I started patient with the understanding I would be terrible. Anyway chord changes were a struggle and the app I am using made a recommendation to get at least 30 chord changes before moving on. This was a struggle and then one Saturday I hit 26 clean chord changes.

I like to be prepared.

Here is a playlist.

https://www.joanwestenberg.com/about/

My two cents- the posts display the sociological imagination. The posts connect the “interplay of man and society and the history of self and world” (p. 4).

https://ratical.org/ratville/AoS/TheSociologicalImagination.pdf

What I found most helpful was to think about the the types of information that we deal with

• Tasks

• Ideas

• Notes

• Resources

These are similar concepts found in other systems notably GTD.

C = Capture

Getting items out of your head and into a system.

O = Organize

Processing what has been captured. Keep track of its state-Not started, In Progress, Completed, etc.

R = Review

Updating the capture list and keeping its status current and identify what needs to be done.

E = Engage

Doing the work.

I asked Claude

What happens if I don’t finish in the first engage cycle?

The cycle starts again.

Capture → Organize → Review → Engage → (new info) → Capture → Organize → Review → Engage

The cycle ends when the task is completed.

This Repo helped me figure out setting up the input values using a data frame including correcting converting the data. The quarto page helped me figure out how to filter the data.

Claude provided a conversion steps from ggplot2 to Observable JS plot.

Yes, you can convert a ggplot2 script to an Observable JS plot. Here’s an approach to translate ggplot2 concepts to Observable JS using the Plot library:

import * as Plot from "@observablehq/plot";

// Assuming similar data structure to R/ggplot2
const data = [
  {category: "A", value: 10},
  {category: "B", value: 15},
  {category: "C", value: 12}
];

// Equivalent to ggplot(data, aes(x=category, y=value)) + geom_bar(stat="identity")
const barPlot = Plot.plot({
  marks: [
    Plot.barY(data, {
      x: "category", 
      y: "value",
      fill: "category"
    })
  ],
  x: {label: "Category"},
  y: {label: "Value"},
  color: {legend: true},
  title: "Bar Plot Conversion"
});

// Scatter plot equivalent
// ggplot(data, aes(x=x, y=y, color=category)) + geom_point()
const scatterPlot = Plot.plot({
  marks: [
    Plot.dot(data, {
      x: "x", 
      y: "y",
      fill: "category"
    })
  ],
  color: {legend: true}
});

// Line plot equivalent
// ggplot(data, aes(x=x, y=y, group=category)) + geom_line()
const linePlot = Plot.plot({
  marks: [
    Plot.line(data, {
      x: "x",
      y: "y",
      stroke: "category"
    })
  ]
});

// Render the plot
document.body.append(barPlot);

Key conversion tips: - ggplot() → Plot.plot() - geom_bar() → Plot.barY() - geom_point() → Plot.dot() - geom_line() → Plot.line() - Aesthetic mappings (aes) become configuration objects - Styling and legends are handled similarly but with different syntax

The main differences are in the configuration approach and method names. Observable JS Plot is more functional and uses method chaining differently from ggplot2’s layered approach.

This posteevie from Seth Godin has some key quotes for me to think about. At first, I thought it was a justification for process culture, but not exactly. There is more nuance in that being 100% focused on process may not produce breakthroughs, but an organization without any process will flame out.

Process is the investment we make in inefficiency now to prevent errors from costing us later.

Godin provides the example of an emergency room; it is process-focused but not very efficient, but the system does not rely on good luck.

I see knowledge work as sophisticated decision making weighing trade offs of unpredictable outcomes judged by irrational humans who disagree with you and each other.

Our agency has to set up other agencies and individual to do their piece of the puzzle well; the manual can’t cover everything and there are often irrational humans involved. When there is disagreement there needs to be negotiation.

Many years ago, I received feedback with the suggestion to change the default message. However, I could not figure out how to make this change.

However, I took another stab at it and was able to figure it out with the help of Claude.ai. I

shinymanager::set_labels(
    language = "en",
    "Please authenticate" = "Please log in"
)

The required change was easy but not obvious to me. I placed this code in the run_app.R function in my Shiny Application

I thought this cant’ be serious and when I previewed my survey; I knew the situation was dire. Thankfully, I recalled Qualtrics has Version history and I was able to revert my survey to a point where all the display logic was working correctly.

Here the general steps to revert a survey back to a working version.

1). While in Survey Builder

2). Go to Tools

3). Select Versions

4.) Find the point where the survey was how you need it.

Once I reverted the survey back to an acceptable state, I added the new display logic. I am now very thankful for the survey version feature.

For me, the gap between the idea and production is speeding up for a Shiny Application that I maintain–thanks to LLMs.

I wanted to improve a section in my application that displayed the count of surveys entered by site. My prompt for Shiny Assistant was not great and got this Toy Application. Did I replicate this? No but it set me in motion on what could be done.

Guidance could be found here at the Quarto Website

I needed to add this to my quarto yaml.

website:
  title: "myblog"
  site-url: https://www.myblogexample.io
  description: "A great sample blog"
  navbar:
    right:
      - icon: rss
        href: index.xml

• Please note I needed to reference blog.xml rather than index.xml

Below needed to be added to main index.qmd file.

---
title: "myblog"
listing:
  contents: posts
  sort: "date desc"
  type: default
  categories: true
  feed: true
---

I have been using Positron and noticed it is a bit different workflow and its confusing because after a preview I can push new content. I noticed my feed was not being created. I went back to Rstudio to rebuild the site. After, I rebuilt in Rstudio my feed was created and I could subscribe to my own content.

My random idea was just two insert a column function in the box; however, when I added the column I received an error. I added two columns but the tables were still stacking on top of each within the row. Clause AI suggested to wrap the two column in a fluidRow statement and it worked like a charm.

 fluidRow(  # Add this fluidRow
      column(
        width = 6,
        h3("Pre Outcome Site Report"),
        gt::gt_output(ns("pre_site_report_23"))
      ),
      
      column(
        width = 6,
        h3("Pre Process Site Report"),
        gt::gt_output(ns("pre_site_process_report_23"))
      )
    )  # Close the fluidRow

## Create lists
df_list <- split(dupes, dupes$response_id)


### Split the list as save a data frame
df_1 <- as.tibble(df_list[1])

df_2 <- as.tibble(df_list[2])

all.equal(df_1, df_2)

Compares the date frame and I could easily tell there were many differences between these two entries. If I had time I would use fake data. Here is the 📦documentation

I was happy this morning that my GHA ran but was surprised that it ran early than expected. I realized that I need to factor in time zones and that my schedule was set for UTC. Claude A.I. was able to help me reset it for Eastern Standard time.

I chose the following and it worked like a charm.

div(
    style = "margin: 20px;",     # Or use padding: 20px;
    # Your plots and tables here
  )

Introduction

SQL is on my mind and this looks like an interesting resource to test those SQL skills. I struggle with new data sets so this may be a good challenge.

I found the resource from this post on bluesky.

Starting my new data analyst job next week and wanted to have a refresh of SQL.

Ran through the SQL Murder Mystery (mystery.knightlab.com) and SQL Island (sql-island.informatik.uni-kl.de).

Nice confidence boost seeing how much quicker and more efficient I was compared to a few months ago

[image or embed]

— Tim (@talbtree.bsky.social) November 4, 2024 at 3:46 PM