Sort the backlinks of one of your note by their number of backlinks to find notes that are important to you.

Every so often I get into the top N reading lists of other people and always find interesting tips. The other day, while sorting my notes, I thought: what is a reading list I would suggest?

The immediate next question was: can I automate that?

The good news is that I take notes for the books I read, so the question became simpler: how can I use my Org Roam notes to rank my readings?

But a more general question is: when is a note interesting? A note seems to be more interesting than another if other notes refer to it. Really I am taking inspiration from the Page Rank algorithm here because it resulted quite successful.

All of this started from creating a reading list, but what about sorting notes for one of my readings? I want to get a possible answer for: what is the single most important thing I learned from this book?

And we can have similar questions for other notes, for example: what is the most important backlink to this note?

All these questions sound similar to me. And I think we can answer them all with the same trick.

Let's say a note is important if it has many backlinks. The rationale is that if we add many backlinks to a note, we spent quite some time on that note. Since time is valuable, a note on which we spent a lot of time is very valuable. But wait! Let's go a step further: if we snowball the approach, we can find how many backlinks each backlink has. That way we find how much time we invested on the notes that link back to the notes that link back to the one we are interested.

(I will stop there because I think I reached my mental capacity with that!)

How does this look in Elisp?

(--> "book"
     org-roam-node-from-title-or-alias
     org-roam-backlinks-get
     (-map (-compose 'org-roam-node-title 'org-roam-backlink-source-node) it)
     (--map (list
             :title it
             :n-backlinks
             (condition-case err
                 (--> it
                      org-roam-node-from-title-or-alias
                      org-roam-backlinks-get
                      (--map (org-roam-backlinks-get (org-roam-backlink-source-node it)) it)
                      (-flatten-n 1 it)
                      (or (ignore-errors (length it)) 0))
               (error
                -1))
             )
            it)
     (--sort (> (plist-get it :n-backlinks)
                (plist-get other :n-backlinks))
             it)
     )

Since I tag my readings with a backlink note titled "book", I start from that note title. Then I collect its backlinks, find the titles of the backlinks and then count the number of backlinks each of these have. Finally sort in descending order by number of backlinks.

This simple approach produced both expected and surprising results: like a book about Scalability was as important to me as a book about real estate!

When I substitute the title of one of the books notes (that is one of the backlink of "book") in the function, I find out which note(s) made that book so important to me.

So, the above code snippet is more useful if wrapped in a function:

(defun ranked-backlinks (note-title)
  (--> note-title
       org-roam-node-from-title-or-alias
       org-roam-backlinks-get
       (-map (-compose 'org-roam-node-title 'org-roam-backlink-source-node) it)
       (--map (list
               :title it
               :n-backlinks
               (condition-case err
                   (--> it
                        org-roam-node-from-title-or-alias
                        org-roam-backlinks-get
                        (--map (org-roam-backlinks-get (org-roam-backlink-source-node it)) it)
                        (-flatten-n 1 it)
                        (or (ignore-errors (length it)) 0))
                 (error
                  -1))
               )
              it)
       (--sort (> (plist-get it :n-backlinks)
                  (plist-get other :n-backlinks))
               it)
       ))

Get some insights from your notes! And when you come up with the your most valuable notes, think about sharing them ;)

Happy ranking!