A browser-based node editor for connectome data. Wire up the steps — query, traverse, filter, aggregate, plot — and each one shows its result where it sits.
Every step is a card. Find Neurons runs one query: match a type pattern
against a dataset. Here, traced lobula columnar cells — ^LC\d+$.
Inputs left, outputs right, parameters between them, the last result along the bottom. ▶ in the header runs this node and whatever is upstream of it.
The stripe down the left edge is run state. Grey here because the Dataset input is empty, so there is nothing to run yet.
A link carries a typed value, not a file. Each socket declares what it accepts, so a link that cannot work is refused as you drag it.
neuronIdColour names the material and shape separates the members within it, so a shape repeats across materials and never inside one: a skeleton and a table of Neurons are both filled circles, Matrix and Transform both diamonds, the hue says which. One material can hold several shapes — the three marks above are all Geometries.
Each node has a cost. Cheap ones — Filter,
Sort, Group By — re-run automatically after you
edited them. Expensive nodes query e.g. a backend and go stale instead.
Run (⇧R) executes everything stale, in dependency order. ▶ on the card itself executes that node and only its upstream. Both report progress as a gold ring round the card.
Auto-run means every edit re-runs the whole graph, expensive nodes included. On by default but can be turned off, so an expensive workflow is under your control.
Clear removes all results, so you can start over. It does not remove any nodes or wiring. Also worth trying if a node is stuck in e.g. an error.
Coda provides nodes for complex operations such as Find Neurons,
Connectivity, Paths or Skeletons, and
primitives for basic data processing steps - e.g. Filter, Group By,
Join, Pivot and so on.
Use the "+" button to add new nodes, or press Tab to open the node browser. Search, category chips, and a port signature on every row, so you can see what a node takes and returns before inserting it.
Space is the same set as a compact palette, plus every command. Drag a wire into empty canvas and the list narrows to what accepts it — then wires up your pick.
Bar Chart draws the table and passes the same table on, so a viewer
drops into the middle of a chain without breaking it. Table,
Heatmap, Scatter Plot, Network Viewer,
3D Viewer and an embedded neuroglancer all behave this way.
Several also emit a selection. Lasso points in Scatter Plot, click nodes in
Network Viewer: the picked rows leave by a Selected output
and wire onward like anything else.
A selection is an ordinary param: it lands in the saved file and in every downstream provenance key. Picking by eye is a step, not a detour out of the pipeline.
A workflow is a simple JSON file that captures every step, every parameter.
Save ▸ Download writes it as a .coda.json.
Save ▸ Save stores it in your browser’s local storage,
so you can close the tab and come back later. Share
A workflow file never contains data/results or your credentials! Whoever opens it gets this pipeline exactly as you left it, presses Run, and can move a threshold to see what changes.
Connectomes are complex datasets. Coda provides a few simple widgets to get started.
Explore Dataset enables fuzzy search
A dataset’s whole neuron table is downloaded once and searched locally. Type something into
the search box and the table filters down to matching rows. Alternatively, use the
simple query language to match on specific columns: type==LC4. Regex also
works: type~^LC[0-9]+$. Select neurons of interest to send them downstream into
your analysis.
Find Neurons for structured searchDefine and combine filters on any column, and the table updates in real time. Where the backend supports it, this will run a live query and return the results. Otherwise, it will filter the local table.
Dataset Summary provides a birds-eye viewThis nodes summarizes the dataset: annotations, neuron counts, completion rate, etc. What is shown depends on what the dataset publishes.
Neuron Profile provides neuron-level summariesIdentity, transmitter, synapse totals, strongest partners, the per‑ROI split and the morphology — one neuron at a time. Page through a collection freely; pin one and it leaves by Current.
Coda does not host any connectomes itself. Instead, we fetch the data from the primary sources, such as neuprint.janelia.org.
Even public connectome datasets often require you to make an account. When trying to access them via Coda, you will be prompted to either log in with the same account or paste an authentication token.
Held in local storage, never sent to a third party or written into a saved workflow.
A .coda.json you send a colleague carries the analysis and not your
credentials.
Some datasets are hosted on multiple backends. For example, the public release of the FlyWire dataset is available via both neuPrint and CAVE. Coda supports both but note that each of these backends has slightly different data and capabilities.
Every published dataset node arrives with a Description card wired to it:
what the dataset covers and the papers its authors ask you to cite. Delete it if you
like but please do not forget to give credit where it is due.
Each type of backend requires a different token. Coda will prompt you the first time you try to access a dataset, and then store it in local storage for future sessions.
Everything can be done with the mouse, but keyboard shortcuts speed up the workflow. For power-users: Space brings up a command palette.
Initials work in the palette: ra → Run All, gb → Group By.
Four places, and one of them has strings attached.
.coda.json filePlain JSON — put it in a repository, a supplement, an email, and reproduces the pipeline wherever it lands. Results are not in it: the recipe, not the meal.
A named copy in this machine, listed under Open and on the start page. Good for picking up where you left off.
Not a backup! Per browser and per profile, not synced, empty in a private window, and cleared with the site data. Anything you would be sorry to lose, download.
The JSON encoded into a shareable link. Paste it into an email or a chat, or just save it in a note for yourself.
A workflow file never contains results. If you want to keep the outputs you have to download them as CSV, SVG, PNG, etc. via the either the Download button in a node's footer, or via the dedicated Download node.
Nothing to install, and the demo data needs no accounts. Start from one, take it apart, point it at your own question.
Coda is in beta. The core is in place; we’ll keep adding nodes and fixing bugs, but a workflow saved today should keep opening in later builds.