Custom chart library for embedded analytics teams

The moment this becomes a real problem: a data or BI team finishes connecting all their data sources, gets sign-off on the dashboard layout, and then discovers they cannot render the chart their stakeholder actually needs — a Sankey diagram, a custom geo map, a bubble chart with more than a few dozen points — without exporting the data and rebuilding the whole thing in a separate tool. That's not a minor annoyance; it's a workflow collapse that adds days and creates a split source of truth.

The gap persists because the major analytics platforms are fundamentally infrastructure businesses. Their engineers are incentivized to improve query performance, connectors, and data governance — not to maintain a sprawling library of niche chart types with deep styling controls. Visualization polish doesn't show up in security audits or procurement checklists, so it rarely gets prioritized. The buyer (IT or data infrastructure) and the user (analyst or product manager) are different people, which means the person who feels the visualization pain isn't the one who signed the contract.

What's actually missing isn't 'more charts' in the abstract — it's a renderer that accepts a data query result from any of these platforms and outputs a fully styled, interactable chart that can be embedded back into the same dashboard or exported as a live iframe. Users specifically call out D3-style interactivity, maps, and the inability to control point density in scatter plots. These are solvable rendering problems, but they require deep front-end investment that a query-first platform has no incentive to make.

This is a business and not a feature because the need recurs every time a new dashboard project starts. Data teams don't hit the visualization ceiling once — they hit it on every project that involves a non-standard chart type. Without something like this, the workaround is to maintain a parallel Tableau or custom D3 environment just for edge-case charts, which means double the licensing, double the maintenance, and data that lives in two places.

What to build

Build a self-hostable chart renderer that accepts a JSON data payload from any analytics platform's query API, renders configurable D3-based chart types (Sankey, geo map, violin plot, dense scatter), and returns an embeddable iframe or web component with live tooltips and drill-through events.

Where to start

Start with Apache Superset users specifically, since Superset is open-source and its plugin architecture is documented — a native Superset plugin that adds five high-demand chart types gets organic traction in the Superset community before you need to sell anything.

The hard part

The hardest early challenge is convincing a team to add a third-party renderer inside a dashboard they already own — the integration story must be zero-friction from day one, meaning the first version needs working embed snippets for at least two major platforms before anyone serious will try it.

How it makes money

Free tier for up to three chart types and self-hosted use; paid plans per seat per month for teams that want the hosted rendering service, additional chart types, and priority styling support.