Importing and exporting

Click Import / Export in the top bar of a map, in any mode. To start a new map from a file, use Import on the home page instead (see Managing maps).

What you can import

Choose one of three shapes, up to 10 MB in total:

Shape	Files	What it brings in
GraphML	One .graphml or .xml file	Nodes and connections, already linked up. Usually exported from another tool.
CSV, nodes only	One .csv file	Nodes only, one per row. Good for getting a list onto the map to connect later.
CSV, nodes and connections	Two .csv files	A file of nodes and a file of connections. Each connection names the nodes at its ends, so both files must come in the same import.

The spreadsheet conventions match Kumu’s, so a spreadsheet prepared for Kumu imports unchanged. The panel’s Which CSV columns to use section has a quick summary. The full rules are in Spreadsheet and GraphML format.

Import step by step

1. Choose your files. Click a shape to pick files, or drag them onto it. The files upload, and the tool reads their columns (“Uploading and scanning…”).
2. Say what each file holds. For a two-file import, the tool suggests which is the file of nodes and which the file of connections, from their columns. Correct it if it’s wrong. You only see this step when there’s something to decide.
3. Map the columns. Each file gets a section with sample rows and one row per column. For each column, confirm or change what it becomes, such as Households → measure, Type → tags (one per value), or Ignore. Options that can’t work for a column aren’t offered.
   • Type a name in the last column to rename what a column creates. For a per-value column, that name becomes the tag group, so a Type column renamed “org type” puts “tenant union”, “funder” and “public body” under “org type”.
   • If the map already has a tag, measure or connection type with the column’s name, the column links to it instead of creating a duplicate.
   • Warnings appear at the top. They include rows that will be skipped, duplicate labels, connections whose ends can’t be found, and columns that look like personal data. They don’t stop the import.
4. Click Import. The import runs (“Importing…”), then shows a summary: how many nodes and connections were created, what was skipped and why, and any notes about direction.

For GraphML, step 3 lists the file’s attributes instead of columns. Each can be Link to existing, Create new, Keep as property, Ignore, or for a node attribute Use as node label. An attribute like “type”, with a few repeated values, is split into one row per value, so each value can become something different.

After the import

Imported nodes go to the Unplaced list in the Nodes card, not onto the canvas. See Placed and unplaced nodes. Connections between them exist straight away, and are drawn once both ends are placed. Visualise mode shows everything without placing it.

Each import creates new nodes. Importing the same file twice gives you two copies of everything. To change a map after importing, edit it in the app or in Table mode.

Uploaded files are stored privately in the EU, and deleted when the map is deleted.

Exporting

In Import / Export, choose a format and click Download. You can also download from a map’s ⋯ menu on the home page.

• GraphML (full graph): every node and connection, with their measures, tags and properties. Undefined connections are left out.
• CSV (nodes only): one row per node, with its properties, measures and tags as columns. It doesn’t include connections, so use GraphML if you need them.