Learn · Obsidian for researchers
Most qualitative software gives you a whiteboard: you drag boxes around and the boxes know nothing. Obsidian Canvas can do something those tools cannot — because every box can be a real note, dragging a code onto the canvas brings its actual quotes and counts with it. The map stops being a picture of your analysis and becomes your analysis, arranged.
After this guide you will be able to:
In a traditional CAQDAS whiteboard, a box labelled "waiting" is a drawing of the word "waiting". It has no connection to the excerpts you coded. If you rename the code or add ten quotes, the box does not change — it is a picture, and pictures go stale.
On an Obsidian canvas, you drop the actual note for that code onto the board. It shows the note's live contents; double-click and you are reading the coded quotes themselves. Rename the code, add excerpts, and the node reflects it, because the node is the note. Nothing is a representation of your data. It is the data, placed where you can see how it connects.
Lay out one column per theme. At the top of each, the theme's note, with a small card above it summarizing its grounding. Beneath, the code notes that feed it, each carrying its own live counts. Attach the negative cases in a different colour, with an edge that says "complicates".
Now the map earns its keep. You can see, without opening anything, that this theme rests on three codes and seven quotes across four participants, that two negative cases push against it, and that a human has signed it. Click any box to check the claim against the real excerpts underneath.
One discipline keeps an evidence map honest: the card shows grounding — how much evidence sits under a theme and whether counter-evidence has been sought — never a score for whether the theme is "true". No number can certify a theme, and a map that displayed one would be inviting you to trust the layout instead of the data. Read "negative cases: none linked" as a task you have not done yet, not a clean bill of health.
An evidence map is most useful when it shows you what is wrong. Three patterns to look for:
Because you rearrange the map by dragging real notes, thinking through these problems is the analysis — you are not annotating a picture of your work, you are moving the work itself.
| Traditional CAQDAS map | Obsidian evidence map | |
|---|---|---|
| The node is | a drawn shape with a label | a live note |
| Click it | nothing, or an editable box | opens the real coded quotes |
| When data changes | the drawing goes stale | the node reflects it |
| Portability | proprietary project file | plain files you own |
The big tools treat visual mapping as a side feature — a whiteboard bolted onto a database. Because every object in Obsidian already exists as a linked Markdown note, the map and the data are the same thing. That is not a cheaper imitation of NVivo's map. It is a different, and for auditability a better, kind of object.
Building it by hand: in Obsidian, create a new Canvas, then drag notes from the file list straight onto it — each becomes a live card. Add a plain text card above a theme for its grounding summary, and colour the negative-case cards to stand apart. It is manual, but it is real: every card you place is a note you can open.
You can assemble an evidence map yourself by dragging your code and theme notes onto a Canvas.
If you want it generated — a map drawn automatically from your reviewed coding, with grounding cards, negative cases attached, and codes not yet in any theme grouped honestly — QualiVahti Local builds exactly this evidence map, and ships a worked demo study so you can see a finished one first.
See QualiVahti Local — €49The set: Obsidian for researchers — 5 questions no guide answers and build a qualitative codebook in Obsidian. More on the Learn page.