Docs/Working with data

Working with data

In Minerva, everything you write is also data you can question. As you work, your notes become a knowledge graph and your tables become a database — built automatically, kept current, and queryable. So instead of only reading your notes back, you can ask precise questions of them and get answers, right inside a note or from a dedicated panel.

Two kinds of data

Minerva keeps your work as data in two forms, each with its own query language:

The knowledge graph
A precise map of your notes and how they connect — titles, tags, links, properties, and any facts you write down. It's built automatically from your writing on every save, and you query it with SPARQL. This is what lets you ask things like "every claim tagged unresolved" or "which notes contradict this one."
Tables
Your CSV files and note tables become typed tables in a built-in database, which you query with SQL — the familiar language for rows and columns. Good for numbers, records, and anything tabular.

What becomes data

You don't set any of this up or maintain a schema — Minerva reads the structure out of ordinary writing. These are the parts of a note that fold into the graph automatically:

Titles & tags
Each note's title and every #tag become things you can filter and group by.
Links
Wiki-links and typed links (supports, rebuts, depends-on…) become the edges of the graph — the connections you can traverse.
Properties
The properties at the top of a note — status, dates, and any fields you add — become structured details attached to it.
Tables & CSV
Markdown tables and .csv files become queryable tables.
Embedded facts
Blocks of Turtle / RDF let you state structured facts by hand, right inside a note, when you want more than tags and links can express.

What you can do with it

Once your work is data, you can query it, compute over it, and visualize it — each has its own page:

🔎

Query menu & panel

A dedicated place to write a question, run it, and reuse it — with ready-made and saved queries.

 ▶️

Runnable cells

A question you write and run right inside a note, with the answer appearing just below it.

 📊

Result blocks

Drop a question's answer into a note as a live list, table, or chart that stays up to date.

 📋

Tables & CSV

Turn CSV files into typed SQL tables you can query and chart.

 🧮

Compute cells

Run Python over your notes and tables, with output rendered inline.

 📈

Charts

Turn query results and tables into interactive charts, live in the note.

 🐢

Turtle / RDF

State structured facts by hand when tags and links aren't enough.
Asking questions yourself vs. asking the assistant

When you chat with the assistant, it can look things up in your knowledge base for you — see Conversations. This section is about doing that yourself, whether from the Query panel or right inside a note.

Where your data lives

Both forms are generated from your files and stored inside the thoughtbase's hidden .minerva folder — the knowledge graph as a Turtle file, the tables in a local database. You never edit these directly: they're rebuilt from your notes as you work, so your markdown files always remain the source of truth. If something ever looks out of date, Maintenance can rebuild them on demand.

Data operations

The Query menu gathers the commands for asking questions and taking the graph out — running a new query, stock and saved queries, and exporting the knowledge graph. See Data operations for each one.