:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.
Python ab-testingai-analyticsanalyticscdpdata-warehouseexperimentsfeature-flagsjavascriptproduct-analyticspythonreactsession-replaysurveystypescriptweb-analytics
import { type ReactElement } from 'react'
import { emptyStateIllustration } from '@posthog/mcp-ui'
import { Empty, EmptyDescription, EmptyHeader, EmptyMedia } from '@posthog/quill'
import {
DefaultTooltip,
type Series,
TimeSeriesLineChart,
type TimeSeriesLineChartConfig,
type TooltipContext,
} from '@posthog/quill-charts'
import { ChartHeader } from './ChartHeader'
import { colorAt, useMcpChartTheme } from './charts/theme'
import type { StickinessVisualizerProps } from './types'
import { getSeriesLabel } from './utils'
const TITLE = 'Stickiness'
const STICKINESS_TOOLTIP_CONFIG = { pinnable: true, placement: 'top' as const }
// The backend returns absolute actor counts per bucket; stickiness reads each bucket as its share
// of the series' total users (`count`). A zero-total series contributes 0% — never a raw count —
// so it can't plot unscaled values onto the percentage axis.
function toPercent(value: number, count: number): number {
return count ? (value / count) * 100 : 0
}
function formatPercent(value: number): string {
return `${value.toFixed(1)}%`
}
const renderStickinessTooltip = (ctx: TooltipContext): ReactElement => (
<DefaultTooltip {...ctx} valueFormatter={(value) => formatPercent(value)} />
)
export function StickinessVisualizer({ query, results }: StickinessVisualizerProps): ReactElement {
const theme = useMcpChartTheme()
const hasData = results?.some((item) => item.count !== 0 && (item.data?.length ?? 0) > 0)
if (!results || results.length === 0 || !hasData) {
return (
<div>
<ChartHeader title={TITLE} />
<Empty>
<EmptyHeader>
<EmptyMedia>{emptyStateIllustration('chart')}</EmptyMedia>
<EmptyDescription>No data available</EmptyDescription>
</EmptyHeader>
</Empty>
</div>
)
}
// X-axis is the interval-count distribution ("1 day", "2 days", …), not a time series. Prefer the
// API's own bucket labels; synthesize "N {interval}s" from `days` when they're absent.
const interval = query?.interval ?? 'day'
const labels =
results[0]?.labels ?? results[0]?.days?.map((d) => `${d} ${interval}${Number(d) === 1 ? '' : 's'}`) ?? []
const series: Series[] = results.map((item, i) => ({
key: String(i),
label: getSeriesLabel(item, i),
data: (item.data ?? []).map((value) => toPercent(value, item.count)),
color: colorAt(i),
}))
const config: TimeSeriesLineChartConfig = {
// No xAxis date config — labels are interval-count buckets, rendered verbatim.
yAxis: { tickFormatter: formatPercent, showGrid: true },
showCrosshair: true,
tooltip: STICKINESS_TOOLTIP_CONFIG,
legend: { show: series.length > 1, position: 'bottom' },
}
return (
<div>
<ChartHeader title={TITLE} />
<div className="flex flex-col w-full h-[400px]">
<TimeSeriesLineChart
series={series}
labels={labels}
theme={theme}
config={config}
tooltip={renderStickinessTooltip}
/>
</div>
</div>
)
}