top of page

Knowledge management to become a sane writer - Part 2

Here I described what the Zettelkasten is and what are its components. Now let me show you how I have built my Zettelkasten in Notion to manage all the information that I come across.

First things first, why Notion?

The quick answer is: it was a personal decision. I could have built my Zettelkasten system with many other apps such as Obsidian or Roam Research. I could have built it in an analogue form. After all, Niklas Luhmann had only pen and paper at his disposal when he created the Zettelkasten method!

The truth was that I was already using Notion, and I was happy with it. Thus, I searched online whether Notion could have been used to build a Zetterlkasten system, or I would have had to invest time in learning a new application more fit for purpose. I realized that I could stick with Notion.

I won't compare here how different apps perform in the creation of Zettelkasten-like systems. It is out of the scope of this post, and the topic is so extensive that it would deserve a (big) post on its own. However, I refer you to some material that I have found interesting about this topic. The main takeaway from that research is that no app is 100% perfect for this task. Thus, the decision remains personal with no right or wrong choice.

Graph of the links between notes in Obsidian.

The only comment I have regards Obsidian and Roam Research's feature that creates a graphical view of the links between notes. Its utility lies in the capability to see at a glance which notes are linked to each other, and what sort of connections have been made. Unfortunately, this feature is not available in Notion. This could be a tipping point for someone abandoning Notion to create a Zettelkasten system. In my case, that was not a big issue.


My Zettelkasten system in Notion

My Zettelkasten system (i.e. inbox for fleeting notes, slip-box for literature notes and slip-box for permanent notes) is built using relational databases. Before I lose you completely, relational databases in Notion are less scary than they sound. Although very powerful, databases in Notion are basically tables. I refer you to this video that explains really well how databases work in Notion (better than I would be capable to do). I used some of the features explained in this video and I will be showing you them in the coming paragraphs.

Only one more note. Notion has recently released new features for their databases. Thus, the pictures in this post might not represent the most up-to-date graphical view of Notion databases. However, the changes are minimal and do not affect the description that I'm doing here of my Zettelkasten system.

My Zettelkasten is contained in one page in Notion, called "Zettelkasten" (yes, I know. I have lots of imagination with names!). The page contains three sub-pages. Each sub-page contains the databases that constitute the "Fleeting Notes Inbox", the "Literature Notes Slip-Box", and the "Permanent Notes Slip-Box". That's it!

Fleeting Notes Inbox

The "Fleeting Notes Inbox" database contains only three properties (i.e. columns): the Note Title, the Creation Date, and the URL (if the note comes from a webpage). I prefer to view this database only as a table and sort the notes with the old ones on top since they are the most urgent notes to sort.

I use two methods to add notes to the "Fleeting Notes Inbox". Directly from Notion, by simply adding a new row to the database. You can easily do this from your computer or your phone with the mobile app. This way to add fleeting notes in the inbox happens when I have an idea, and I want to capture it immediately to not forget it. This is the simplest way to add fleeting notes to the inbox, but it is not the most exciting one.

What became a game-changer for me was the discovery of Notion's web clipper. As a PhD student, I need to integrate data from many sources. My primary reference is still peer-reviewed articles, of course. But today, we are exposed to a much wider range of repositories of cutting-edge ideas, such as blog posts or Twitter! Not always, but looking for a blog post that caught my attention can be relatively easy. However, have you ever tried to find a tweet that contained an interesting idea or pointed to an interesting paper? I have! And 9 times out of 10, I don't find it again. Not surprising since the lifetime of a tweet is of the order of magnitude of a few hours. Therefore, it is vital for me to be able to store that tweet somewhere easy to find at a later stage when I have time to follow it up.

When the note is opened in Notion, all the properties are automatically filled out. In addition, Notion will create a clickable preview of the tweet that can direct you back to the original one on Twitter.

Another cool functionality is provided by the mobile app, which lets you add to the "Fleeting Notes Inbox" whatever type of content (e.g. web pages, photos, videos, etc.) directly from your phone. You just need to click the share button and select the Notion icon on the options that appear on your phone.

There are other systems to synchronize notes such as Readwise or Instapaper. However, by using the Notion web clipper I don't feel the need to use third-party applications.

Literature Notes Slip-Box

The "Literature Notes Slip-Box" database contains seven attributes: Note Title, Status, Related to Perm Notes, Creation Date, Reference, Page, and Relevance. The Note Title is the first tool that will allow me to find the note when writing. For this reason, I tend to write them as questions (so the content of the literature note contains the answer). The Status allows me to identify at a glance which notes have been already "Processed" or are "Under processing", thanks also to the colour code assigned to the different status property. The Related to Perm Notes allows me to see which permanent notes link to specific literature notes. The Reference and Page attributes contain important metadata about the original source of the literature note. This allows me to track back the original source of the literature note if I need to revise it. Every note is taken to be used somehow in my research. However, I might not be sure how I might use that note at the time of its creation. The notes that have the value "important" in the Relevance attribute correspond to those notes that I'm currently using in my writing. Then, I can use this attribute to filter such "important" notes and have only those in front of me when I'm writing my papers, instead of an overwhelmingly long list.

I prefer to view the "Literature Notes Slip-Box" in two different ways: as a board or as a table. The board view allows me to see, at a glance, which notes are "Under processing" or are already "Processed". The table view allows me to see all the literature notes ("Under processing" and "Processed") and sort them according to specific needs (e.g. by author, creation date, or reference).

As we showed in Part 1, literature notes must be atomic, i.e., only one concept must be represented. Therefore, if you find yourself writing two or more paragraphs in one note, check whether you are still writing about one single idea. As a rule of thumb, I tend to not have more than a paragraph in each note, and typically notes will be only a few lines. I also like to add comments in the notes, mainly to remind myself what made me take that literature note or how I might use the note's content in the feature. Finally, note that literature notes can also have links to other literature notes. This will be useful at the moment to write your permanent notes, but we'll discuss that later.

Of course, there can be exceptions. For example, the literature note I have about the "Global Precipitation Climatology Project (GPCP)" dataset. It does not follow the rule of "keep it short or under a paragraph". However, the content on this note about the GPCP dataset comes from one paper, so it makes sense to put all this information in one literature note. Furthermore, the note contains also a picture. This shows that adding different types of content (not only text) can be useful. Add whatever you might need to make the literature notes useful when using them to write your permanent notes, but keep them always atomic!

Permanent Notes Slip-Box

The "Permanent Notes Slip-Box" database contains five properties (i.e. columns): Main Topic Tag, Sub-Topic, Sub-Topic Description, Sub-Topic Tags, and Primary Sources. The Main Topic Tag categorises the permanent notes. For example, if the note refers to observations, forecasting, NWP modelling, the note will have tags that recall those general topics. The Sub-Topic attribute corresponds to the title of the permanent note. For example, I would have a note called "Characterisation of flash floods" that summarises the different definitions of the flash flood hazard that can be found in the literature. The property Sub-Topic Description provides a brief description of the content of the permanent note. Permanent notes can (and are actually supposed to) be updated over time. The Sub-Topic Description attribute helps me remember what each permanent note is about in a few words. Like Main Topic Tag allows to sort between the main general topics contained in the "Permanent Notes Slip-Box", Sub-Topic Tags helps to filter the content of the database but in a much more granular way. Finally, Primary Sources lists all the literature notes that build the correspondent permanent note.

Through the Sub-Topic attribute, one can access the content of the permanent note as a Notion page. The note is organized as follows:

As you can see, I added in each permanent note the "Literature Note Slip-Box" that shows only the literature notes that build that specific permanent note. This might seem redundant at first sight since the same information is presented in the summary area of the database attributes. However, having the "Literature Notes Slip-Box" inside the permanent note allows me to add literature notes as the permanent note is written without leaving the permanent note page, which is a time-saver. The literature note will also be automatically assigned to that permanent note, which means the main attributes of that particular literature note will be automatically filled out, saving me time again because I will not need to do it manually.


Conclusive notes

I think it is worth stressing the following. There is no right or wrong way to create a personal knowledge management system. Why? Because as the name says, the system is "personal", and it needs to fit the individual's specific requirements. By looking at other people's systems, one can quickly get excited about recreating someone else's beautiful, complex system. However, such a system might not suit the individual's needs and might end up killing productivity. Indeed, one might spend lots of time creating a perfect, pretty system with different levels of complexity to end up realising that there was no need for such a sophisticated system, and such time might have been better spent using a much simpler system. This is a real problem when using apps like Notion that allows endless customization.

If you are interested in building your own personal knowledge management system, get inspiration from other's systems, but always seek efficiency over aesthetics and complexity. For example, I needed a quick and "dirty" but efficient way to put all my notes in one place and make them easily discoverable during the writing of my papers. So my system does not contain fancy dashboards or complicated automation (at the end of the day, my system is made of three simple databases). It might not be pretty either, but it works just fine for my needs.

Furthermore, don't allow perfectionism to stop you. Your system will never reach perfection as there will always be room for improvement. And especially at the beginning, the system might appear very basic and rudimental. However, knowing that the system is not perfect should not stop you from building it and using it. It will improve over time.


bottom of page