Learn · Obsidian for researchers

Obsidian for researchers: 5 questions no guide answers

Researchers hit the same structural dilemmas with Obsidian, and most guides answer them for note-taking hobbyists, not for people managing literature, data, and a thesis. Here are direct answers to the five questions researchers actually ask — one vault or many, how to handle folders and tags, where Zotero stops, whether to write a dissertation in it, and how to read the graph view for gaps.

1. One massive vault, or split by project?

Use one single vault.

Splitting vaults breaks your internal links. The whole value of Obsidian for research is that a concept from a paper you read in 2024 can surface, unbidden, next to a draft you are writing today — because they link to the same note. Separate vaults cannot see each other, so you lose exactly the cross-project connections that make the tool worth using.

The exception: create a separate vault only for strictly client- or patient-confidential data, or when you are actively collaborating with a team through a shared repository. Isolation there is a feature, not a cost.

2. Folders, tags, and links without making a mess?

Links for ideas, folders for access and file type, tags only for status.

• Links are your primary connective tissue. Link concepts directly inside your sentences — [[neuroplasticity]] — to build an organic web that grows as you write.
• Folders stay minimal. Use them for file types or processing stages: /Sources, /Drafts, /Templates. Not for topics.
• Tags are for workflow status only: #to-read, #needs-citation. Topic tags like #psychology multiply until they are noise; links do that job better.

The rule of thumb: if you would ever want to see the connection, use a link. If you just need to find the file, a folder is enough. If it describes where the note is in your process, a tag.

3. Where does Zotero end and Obsidian begin?

Zotero stores the data; Obsidian synthesizes the thinking.

Keep them in their lanes and neither fights the other:

• Zotero captures PDFs, holds complete metadata, generates your final bibliography, and is where you do raw, colour-coded highlighting as you read.
• Obsidian is where you write notes in your own words, break arguments into atomic concepts, and cross-reference several papers against each other. Use the Zotero Integration plugin to pull your highlights across.

A working line: highlighting is capture, and belongs in Zotero; putting the idea in your own words is thinking, and belongs in Obsidian. If you are copying a quote, that is Zotero's job; if you are arguing with it, that is Obsidian's. New to Zotero itself? Start with our Zotero for researchers guide.

4. Is it safe to write a whole thesis in Obsidian?

Yes for drafting and outlining; no for final formatting.

Draft freely. Brainstorm structure on the Canvas, collect your references, and assemble chapters with a plugin like Longform. Because Obsidian stores everything as plain, standard Markdown text files, your work stays portable — no proprietary format can trap it, plain text opens in any editor today, and it is the format most likely to still open in twenty years.

The export: when it is time for formal layout, style checks, and peer-review tracking, export the raw text via Pandoc to LaTeX or Word. Obsidian is where the thesis is thought and written; those tools are where it is dressed for submission. Keeping the two stages separate is a strength — you never fight a page-layout tool while you are still thinking.

5. How do I use the graph view to find research gaps?

Look for isolated islands, sparse hubs, and long bridges.

The graph view is not decoration; read it as a map of your own thinking:

• Isolated islands — notes floating alone with no links. These are papers you read but never integrated into your core argument. Connect them or ask why they resist connection.
• Sparse hubs — a major concept note with very few papers attached. That thinness reveals either a gap in the literature or an area you have not read deeply enough yet.
• Long bridges — two dense clusters joined by only one or two links. Those thin connections are where original, cross-disciplinary work lives: you are one of the few people standing between two fields.

Qualitative analysts can push this further — instead of the whole vault, filter the graph to your codes and see which ones travel together and which stand alone. That is the idea behind an evidence map.

Putting it together: a minimal research vault

If you are starting from an empty vault, this is the smallest structure that follows all five answers and scales to a whole PhD:

• Three folders, no more to begin with: /Sources (one note per paper, with its Zotero highlights imported), /Notes (your own atomic concept notes), /Drafts (chapters and manuscripts). Add /Templates once you repeat a note shape.
• Two or three status tags, and no topic tags: #to-read, #needs-citation, maybe #revisit. That is the whole tag system.
• Links from day one: every time you write a concept in a note, wrap it in [[brackets]]. The web builds itself as you work, and the graph view starts telling you things within a few weeks.

Everything else — folders for sub-topics, elaborate tag hierarchies, nested categories — is structure you can add later if a real need appears. In practice the need rarely does. The links carry the load.

Read next

Practical setup: build a qualitative codebook in Obsidian, the plugins and features worth adding, and use Obsidian Canvas as an evidence map. More on the Learn page.