From bb27118f3c6e25d301c5c157dcf666e2e47ceee5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ma=C3=ABlle=20Salmon?= Date: Fri, 6 Jun 2025 07:24:35 +0200 Subject: [PATCH 1/4] typo fix only in the qmd --- content/blog/2025/06-01_visible_pca/index.qmd | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/content/blog/2025/06-01_visible_pca/index.qmd b/content/blog/2025/06-01_visible_pca/index.qmd index 1cefa3434..3cdffb415 100644 --- a/content/blog/2025/06-01_visible_pca/index.qmd +++ b/content/blog/2025/06-01_visible_pca/index.qmd @@ -84,7 +84,7 @@ From the analysis, four main clusters emerged, each telling its own story about * *My initial thought:* "This one is a no brainer for me, yeah, that is my brain doing its thing. As a person who generally struggles with depression and low mood, this clustering is meaningful. Biggest surprise here was that anxiety was not clustered in with it all." 4. **Neurological:** This was the largest cluster, including "fatigue," "sleep," "joint pain," "headache," "palpitations," "resting heart rate," "anxiety," and "crash (PEM)." - * *My initial thought:* "This really encapsulates the 'Long Covid feeling.' All these symptoms are the heavy hitters that define my worst days. It's the central hub of what pulls me down. While there is clear meaning behind anxienty being in here (because on poor days I get feelings of despair around not improving), I still though it would cluster together with depression." + * *My initial thought:* "This really encapsulates the 'Long Covid feeling.' All these symptoms are the heavy hitters that define my worst days. It's the central hub of what pulls me down. While there is clear meaning behind anxiety being in here (because on poor days I get feelings of despair around not improving), I still though it would cluster together with depression." From 434a51080b03961a02df77ae8ccf7a57123783a5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ma=C3=ABlle=20Salmon?= Date: Fri, 6 Jun 2025 07:26:17 +0200 Subject: [PATCH 2/4] Update index.qmd --- content/blog/2025/06-01_visible_pca/index.qmd | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/content/blog/2025/06-01_visible_pca/index.qmd b/content/blog/2025/06-01_visible_pca/index.qmd index 3cdffb415..f4bf91f99 100644 --- a/content/blog/2025/06-01_visible_pca/index.qmd +++ b/content/blog/2025/06-01_visible_pca/index.qmd @@ -143,7 +143,7 @@ We look for an "elbow" in the plot, where the explained variance sharply drops o First, we need to create a data.frame we can use for plotting. -For data from prcomp, I like to extract it's summary table and coerce it into a data.frame. +For data from prcomp, I like to extract its summary table and coerce it into a data.frame. It's a little sneaky, but also convenient. ```{r} From 33f57fa794f0deb3b6de8fda001f050b8e24b555 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ma=C3=ABlle=20Salmon?= Date: Fri, 6 Jun 2025 07:27:25 +0200 Subject: [PATCH 3/4] Update index.qmd --- content/blog/2025/06-01_visible_pca/index.qmd | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/content/blog/2025/06-01_visible_pca/index.qmd b/content/blog/2025/06-01_visible_pca/index.qmd index f4bf91f99..de0dbd12c 100644 --- a/content/blog/2025/06-01_visible_pca/index.qmd +++ b/content/blog/2025/06-01_visible_pca/index.qmd @@ -278,7 +278,7 @@ loadings_df |> ) ``` -This is starting to looks much easier to interpret now. +This is starting to look much easier to interpret now. Now, we can focus on the trackers that contribute most on each end of the scale. I'm still searching for a better visualisation though, and I think we can keep the colours, but use their absolute loading value to plot. This will place all the most important components in order at the very left of the plot, and we can distinguish between positive and negative by their colour. From 80a0503597c8dacfcbb66b1f387b38b546c7746f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ma=C3=ABlle=20Salmon?= Date: Fri, 6 Jun 2025 07:28:53 +0200 Subject: [PATCH 4/4] Update index.qmd --- content/blog/2025/06-01_visible_pca/index.qmd | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/content/blog/2025/06-01_visible_pca/index.qmd b/content/blog/2025/06-01_visible_pca/index.qmd index de0dbd12c..59be08830 100644 --- a/content/blog/2025/06-01_visible_pca/index.qmd +++ b/content/blog/2025/06-01_visible_pca/index.qmd @@ -412,7 +412,7 @@ clusters_df <- list( clusters_df ``` -Then we combine the clusters data with the loadsings for the three top components, and the loadings of the factors over `0.2` to keep it neat (else the sankey plot will show lines for every tracker out of each Principal component). +Then we combine the clusters data with the loadings for the three top components, and the loadings of the factors over `0.2` to keep it neat (else the sankey plot will show lines for every tracker out of each Principal component). I'm using the [ggsankey](https://github.com/davidsjoberg/ggsankey) package by [David Sjoberg](David Sjoberg) here, which is _not_ on CRAN, but I think makes the most beautiful Sankey diagram (hint hint David, get it on CRAN, will ya?).