Product Scope Charts
These started as the charts inside one of my own apps. I pulled them out, swapped in made-up data and published them under the MIT licence so anyone building a dashboard, report or admin screen can reuse them without taking on a charting library. The source is on GitHub.
About the charts
Each chart is a JavaScript function that takes arrays of numbers and a few options and draws plain SVG. The toolkit code is one stylesheet and one script, with no dependencies and no external requests, so it runs under a strict content security policy. Colours, font and corner radius come from CSS variables, so it follows your theme, light or dark.
Every chart page gives you when to use it, a live demo, the options and a snippet to paste. The charts handle empty data, a single data point and outliers, can be used from the keyboard and can be shown as a table.
The dashboard is a worked example: ten of the twenty-four charts on one page, wired to made-up data. Claude built it, and the charts. The function signatures are small and regular, so the same approach works for your own dashboards: give an AI assistant the toolkit, describe your data and ask it to build the page.
All twenty-four charts
Each page has a live demo you can change, a copyable snippet and the options. Jump to a group or scroll through them all.
Over time
See how a measure moves, repeats or clusters across days, weeks and months.
Line
One or a few series over time, with a typical-value line and an optional comparison.
View chart →Stacked area
Volume over time, split into themes, or the same data as a share of the total.
View chart →Sparkline
A tiny trend with no axes, for tables, tiles and lists.
View chart →Calendar
One cell per day, laid out in weeks, to show rhythm and one-off spikes.
View chart →Heatmap
A grid of rows against columns, shaded by value, with clear no-data cells.
View chart →Small multiples
One mini chart per category on a shared layout, with a detail chart for the one you pick.
View chart →Comparing
Set one thing against another, in order or step by step.
Grouped bars
Compare two measures side by side per period, with a derived line on top.
View chart →Ranked bars
A top-N list with a bar per row, a detail line and optional click-through.
View chart →Waterfall
How a starting number becomes an ending number, step by step.
View chart →Diverging bars
What went up and what went down: bars either side of a zero line, biggest rise at the top.
View chart →Slope chart
Two moments side by side, one line per item, so rises and falls read at a glance.
View chart →Bump chart
Rank over time: who is first this week, and how each one got there.
View chart →Spread and relationships
See how values are distributed and whether two measures move together.
Histogram
How a measure is spread: thread length, thread time or cost per response, in bars.
View chart →Box plot
Compare the spread of a measure across groups: median, quartiles, whiskers and outliers.
View chart →Scatter and bubble
Two measures against each other, one dot per item, with size, groups and a trend line.
View chart →Parts of a whole
Show what a total is made of and where it goes.
Stacked
Parts of a whole per period, one stable colour per category.
View chart →Donut
Share of a total, with the hovered slice read out in the centre.
View chart →Treemap
Area proportional to value, packed into a rectangle.
View chart →Flow
Where money comes from and where it goes, as proportional ribbons.
View chart →Funnel
A journey where each step can only lose people, with the drop-off between steps.
View chart →At a glance
Single figures, targets, the words people use and the exact numbers behind them.
Gauge
Progress against a budget or limit, with the state spelled out in words.
View chart →KPI tiles
Headline numbers with a change and a sparkline, plus a loading skeleton.
View chart →Word cloud
The words people use most, sized by how often they appear, with optional colour for rising and falling.
View chart →Data table
The accessible twin of every chart: plain numbers in a real table.
View chart →Questions people ask
There is nothing to install and nothing to keep updating, and it takes three steps. First, copy the code from the Get the code page into your page: the design tokens, a few small helpers, one block of CSS and one script. Second, call a chart function such as lineChart with your data and add the result to the page; every chart page has a snippet to start from. Third, make it yours by pointing the design tokens at your own theme, or by asking an AI assistant to build the dashboard for you. If you need something other than pounds, pass your own number formatters.
What are these charts for?
They are for the dashboards, reports and admin screens inside a product: how a number moves over time, how a total breaks down, what needs attention and what to do next. Each chart page says when to use that chart and when not to.
Can I use them to build a dashboard with an AI assistant?
Yes, and that is how this site was made. Give an assistant such as Claude the toolkit page, describe what the dashboard should show and where the data comes from, and ask it to build the page with these charts. Each chart takes plain arrays and a few options, so it can connect them to your own data and restyle them with the design tokens. Check what it produces, as you would with any code, and test it with a keyboard and a screen reader.
Can I use it commercially?
Yes. The code is open source under the MIT licence, so you can use, copy and change it in your own projects, including commercial ones. Keep the licence notice with the code. The fonts on this site are under the SIL Open Font License.
How do I change the currency?
Currency defaults to pounds, in compact(), money0 and gbp. Most charts take yfmt, tipfmt or fmt options to format numbers your way. Calendar, treemap, gauge, waterfall and flow have no such option, so change those three helpers for another currency.
Is it accessible?
Every chart can be reached with the keyboard, has a screen-reader label and can be shown as a data table. There is an option to use patterns instead of colour. If you find a problem, please open an issue on GitHub.
Was this built with AI?
Yes. Claude built all twenty-four charts and the mock-up dashboard, working from my direction. It is the same AI-assisted approach I use on products and reporting for clients, such as the Jira time reporting dashboard I built with Claude.
Want charts or dashboards like these in your own product?
Tell me what you are trying to see and I can help work out the right approach, from the data behind it to the screen in front of people.