Notion vs Obsidian // Databases vs Dataview

Notion and Obsidian diverge at the moment a user creates a note: Notion pushes a top-down hierarchy where every page must be placed in the “right” folder up front, while Obsidian supports a bottom-up network where notes can be linked and tagged first, organized later. That difference matters because it determines whether a knowledge system stays rigidly categorized—or evolves as interests shift and new connections emerge.

In Notion, the workflow starts with a placement decision. Notes live inside folders, and the structure encourages users to begin from broad categories and drill down. The transcript describes how this quickly becomes messy as the number of pages grows: a “YouTube” page might belong under “Content Consumed,” but if the user later starts a YouTube channel, the same topic also fits under “Personal,” and potentially “Work” when content overlaps. The key constraint is that a single page can’t sit in multiple folders at once, forcing users to choose where it “belongs” before they’ve fully understood how they’ll use it.

Obsidian takes the opposite approach. A note can be created immediately without committing to a folder taxonomy. Instead, the system relies on links and tags so one note can be discoverable through multiple paths—by topic, by project, or by any other label that becomes relevant later. The transcript illustrates this with a new note about “this video,” which can be linked to a “YouTube” page and also to “Personal Knowledge Management” without moving it between categories. Tags add another layer: a tag pane can list all notes carrying a tag like “myvideo,” and tags can be stored either in the note body or as front matter.

That same structural contrast shows up in databases. Notion’s databases require upfront decisions: create separate databases for “People,” “Books,” and “TV Shows,” then populate items into the chosen containers. The problem appears when access patterns change. If someone later wants a unified list of “Things I did in 2021” across movies, books, and travel, Notion struggles because items can’t easily be moved into a new table without removing them from the original database. The transcript frames this as a planning trap: the system demands that users predict future queries before they create the database structure.

Obsidian’s equivalent is not a traditional database at all. With the community plugin Obsidian Dataview, the vault’s notes are treated as the underlying dataset, and “databases” are effectively built as queries. Metadata is added to individual notes as needed—fields like type (person, country, etc.) and date (such as 2021). Then Dataview queries generate tables that filter the same underlying notes into different views, such as a “People” table or a “Things I did in 2021” table. Because these views are queries, they can change over time as metadata gets added.

The takeaway is a fit-for-purpose recommendation. Notion suits users who already know the structure they want—work documentation, small wikis, or straightforward task/content lists. Obsidian suits users seeking an open-ended system that can accommodate “chaotic complexity,” supporting serendipitous connections as interests evolve. Dataview is positioned as the bridge that turns Obsidian’s freeform network into something database-like without locking users into categories too early.