HG Insights|Phoenix

Skill: Time-Series Plotting

Time-series rendered in the right shape for the audience — readable at a glance, not at a 30-row table.

Overview

Render time-series data in the shape that fits the audience. Claude picks a sparkline when the trend matters more than the values, a compact table when the values matter, and a labeled chart for multi-series comparisons — and always uses month or quarter axes (not unix timestamps) so a non-technical reader can read the chart at a glance.

Use cases

  • Intent trend in a sparkline, not a 12-row table

    When a brief shows 'intent score over the last 6 months', the output is a one-line sparkline annotated with the metric and window. The reader sees the trend in a second instead of parsing dates and values manually.

  • Multi-series comparisons that actually compare

    Two-vendor or two-account trends render as a labeled chart with units on every axis — not a wall of numbers asking the reader to do the comparison in their head.

View full skill

Time-Series Plotting

When to use

  • Authoring a workflow output that includes time-series data (intent over time, install density over time, headcount over time).
  • Reviewing a generated chart that's the wrong shape for its audience.

Shape decision rule

Data shapeBest representation
3-12 points, single series, trend mattersSparkline (▁▂▃▅▇)
3-12 points, single series, exact values matterCompact table (date / value)
12+ points, single seriesMarkdown line chart (image or rendered)
Multiple seriesMulti-line chart (rendered)
Comparing 2-3 specific datesInline prose ("intent rose from 12 to 87 between Jan and May")

Sparkline conventions

When the shape of the trend matters more than the exact values:

Intent score, 6 months: ▁▂▄▅▇█

Use the Unicode block characters ▁▂▃▄▅▆▇█. Always prefix with the metric name + the time window so the reader has context. Don't use sparklines without labels.

Table conventions

When the exact values matter (e.g., for an audit or a follow-up question):

DateIntent score
Jan 202612
Feb 202628
Mar 202645
Apr 202662
May 202687

Prefer ISO month format ("Jan 2026"); avoid "1/26" style which ambiguates day vs. month.

Axis labeling for non-technical readers

A "non-technical reader" includes most CROs and AEs. Conventions:

  • Time axis: months or quarters, never seconds or unix timestamps.
  • Value axis: include units. "Intent score" alone is fine because the unit is implied (0-100); "Spend" needs "(USD millions)".
  • Annotations: when a known event drove a spike (acquisition, leadership change, product launch), annotate inline — "Apr 2026 spike: new CIO took over".

Common pitfalls

  1. Rendering a 30-row table when a sparkline would do. The table buries the trend.
  2. Rendering a sparkline when the reader needs to act on a specific value. "Did intent hit 70?" can't be answered from ▁▂▄▅▇█.
  3. Unitless axes. "Spend went from 12 to 87" — twelve what?

Reference