ur-grue Derived Metrics Calculator
What This Skill Does
Takes raw numbers you provide and calculates derived figures — percentage changes, per-capita rates, index values, or ratios — with plain-language explanations ready to drop into a story.
When To Use This Skill
- You have two raw figures and need the percentage change between them (e.g. year-on-year budget shifts)
- You need to convert an absolute number into a rate per 100,000 people for meaningful comparison
- You want to express a ratio or share in a form that is immediately understandable to a general audience
- You have multiple data points and want to know which shows the sharpest change relative to its baseline
- A source has given you raw totals but your editor asks for a "per person" or "per household" figure
What You Need To Provide
Required: The raw numbers you want to work with, clearly labelled (e.g. "hospital beds 2019: 42,000 / 2023: 38,500 / population: 5.2 million") Optional: The unit of analysis you want the rate expressed in (per 1,000, per 100,000, etc.); the geographic or time scope; the type of derived metric you want (percentage change, per-capita rate, share of total, index); any known data caveats you want flagged
How the Assistant Approaches This
- Reads the raw figures and identifies which type of calculation is most useful and most defensible for the journalistic context
- Performs the calculation step by step, showing workings so you can check them before publishing
- Rounds to an appropriate number of significant figures for a general audience (usually 1–2 decimal places)
- Writes one or two ready-to-use sentences expressing the result in plain language suitable for a news article
- Flags any caveats — rounding effects, population estimate uncertainty, or comparability issues between time periods — that should be disclosed or checked with a source
Output Format
Short structured response: labelled calculation block showing the arithmetic; one or two plain-language sentences for direct use in copy; a brief caveats note (2–4 bullet points). Total length: under 300 words. Tone is clear and precise — written for a journalist, readable by any audience.
Quality Criteria
- Every step of the arithmetic is shown and checkable without specialist knowledge
- The plain-language sentence matches the numeric result exactly — no rounding or framing inconsistency
- The metric type chosen (rate, change, share) is appropriate to the data provided
- At least one caveat is noted where the data or calculation has a known limitation
- Output is usable in a story without further reformatting
Example
Input
Reported road fatalities in a mid-sized country: 3,840 in 2019 and 2,970 in 2023. Population in 2023: 8.6 million. I need the percentage change between the two years, and the 2023 fatality rate per 100,000 people.
Output
Percentage change, 2019 to 2023
Formula: ((2,970 − 3,840) ÷ 3,840) × 100 = (−870 ÷ 3,840) × 100 = −22.7%
Ready-to-use sentence:
Road deaths fell by 22.7 percent between 2019 and 2023, from 3,840 fatalities to 2,970.
2023 fatality rate per 100,000 people
Formula: (2,970 ÷ 8,600,000) × 100,000 = 34.5 per 100,000
Ready-to-use sentence:
In 2023, the country recorded 34.5 road deaths per 100,000 people — a figure that allows direct comparison with international safety benchmarks.
Caveats
- Population figure is the reported 2023 estimate; if the actual figure differs, re-run the per-capita calculation
- The percentage change does not account for population growth between 2019 and 2023; a population-adjusted rate comparison would be more rigorous for a long-form piece
- Confirm with the source agency whether the 2019 and 2023 counts use the same fatality definition (e.g. deaths within 30 days vs. at scene)
Known Limitations
- This skill performs calculations on numbers you supply — it cannot retrieve or verify the source figures itself; always confirm raw data with the originating agency before publishing
- For complex statistical operations (regression, confidence intervals, significance testing) use a data journalist or statistical software; this skill handles arithmetic-level derived metrics only
- Percentage-change calculations can be misleading when baseline numbers are very small; the output will flag this but the editorial judgement call is yours