AI Prompts for Perplexity Prompts for Statistics and Data

I need the current figure for [SPECIFIC METRIC: E.G. SHARE OF E-COMMERCE IN RETAIL SALES / AVERAGE PODCAST LISTENING TIME / SMB FAILURE RATE]. Find it with full provenance: the exact number, the original source that measured it (not sites quoting it), when it was measured, the methodology in one sentence, and the direct link. If multiple credible sources give different figures, show each with its methodology difference.

Find the market size and growth data for [MARKET / INDUSTRY]: current market value, historical growth rate, projected growth with the projection source and its track record, and how the market is segmented. Market size figures vary wildly by definition: state what each source includes and excludes, and which definition best matches what I mean by [YOUR DEFINITION OF THE MARKET].

What official or authoritative data exists on [TOPIC]? Search government statistical agencies, central banks, international organizations (World Bank, OECD, WHO, Eurostat), regulators, and major academic datasets. For each relevant dataset: what it measures, how current it is, its known limitations, and how to access it. Official data beats industry PR surveys for credibility in anything I publish.

I am writing [CONTENT: ARTICLE / REPORT / PRESENTATION] on [TOPIC] and need supporting statistics for these claims: [LIST THE POINTS NEEDING NUMBERS]. For each: find the strongest current statistic, its original source and date, and the correct attribution phrasing. Where no solid number exists, say so plainly rather than offering a weak or outdated one, and suggest how to reframe the claim without it.

Find how [METRIC] has changed over time: the figures for [TIMEFRAME: EACH YEAR OF THE LAST DECADE / QUARTERLY SINCE 2020], from a consistent source so the numbers are comparable, with the source named. Flag any breaks in the series (methodology changes, redefinitions) that make raw comparison misleading. I need the trajectory, not just the latest snapshot.

Trace this widely cited statistic to its origin: "[THE STATISTIC AS COMMONLY QUOTED]". Find where it actually comes from: the original study or measurement, what the original actually said (often different from the viral version), how old it is, and whether the original source or later research has corrected or retracted it. Many famous statistics are zombie numbers: dead at the source but still circulating.

Verify [STATISTIC / CLAIM WITH NUMBER] across independent sources: find at least three sources that measured this independently (not three sites quoting the same origin), compare their figures, and explain any differences through methodology, timeframe, or definition. Conclude: what is the defensible number or range to cite, and which single source is strongest?

Assess this survey-based statistic before I cite it: [STATISTIC AND ITS SOURCE]. Investigate: who commissioned and who conducted it, the sample size and how respondents were recruited, whether the questions are published (loaded questions produce headline numbers), the response rate, and whether the sponsor benefits from the result. Vendor surveys marketed as research are the most common source of misleading statistics.

Sanity-check this statistic: [STATISTIC]. Does it survive basic plausibility math: what would it imply if true (scale it against known totals like population, market size, hours in a day), does it contradict other established figures, and has anyone credible publicly challenged it? Show the arithmetic. Numbers that fail the envelope math are wrong no matter how often they are cited.

Audit every number in this draft before publication: [PASTE DRAFT / LIST THE NUMBERS WITH THEIR CLAIMED SOURCES]. For each: is it current, does the cited source actually contain it, is it quoted at the right precision and context, and is the strongest available source cited rather than a secondhand one? Flag every number that fails, with the corrected figure and source. My credibility rides on these.

Help me interpret [STATISTIC] correctly: what exactly was measured and what was not, what is the relevant baseline or comparison that gives it meaning (a number alone means nothing), what alternative explanations exist for it besides the obvious story, and what conclusions does it genuinely support versus conclusions people commonly stretch it to support? Search for how experts in [FIELD] contextualize this figure.

The claim: [X IS ASSOCIATED WITH Y, WITH THE STATISTIC]. Investigate the causal story: does credible research establish causation or only correlation, what confounders could produce this association, has anyone run controlled or natural experiments on it, and what do the original researchers themselves say about causality (often more cautious than the coverage)? Give me the honest version of this claim I can defend.

Decode this percentage claim: [CLAIM: E.G. RISK INCREASED 50% / PRICES ROSE 20% / ENGAGEMENT DOUBLED]. Find: the absolute numbers behind it (a 50% rise from 2 to 3 per 10,000 is not what readers picture), the base period and whether it was cherry-picked, and whether relative change is hiding a trivial absolute change or vice versa. Rewrite the claim in the honest form with both relative and absolute framing.

The statistic says the average [METRIC] is [VALUE]. Dig into the distribution: is this mean or median and how different are they (skew means the average describes almost nobody), what do the quartiles or percentiles look like, and are there distinct subgroups being blended into one misleading number? Find the distribution data if published. For [MY PURPOSE], which single number honestly represents the typical case?

This claim needs its denominator checked: [CLAIM: E.G. MOST ACCIDENTS HAPPEN NEAR HOME / BRAND X HAS THE MOST COMPLAINTS]. What is the exposure base that should normalize it (miles driven near home, units sold), does the pattern survive normalization, and is this a case of survivorship or selection bias where what got counted differs from what happened? Recompute the honest comparison if the data exists.

For each statistic I am using, write the attribution phrasing that is accurate and protects my credibility: [LIST STATS WITH SOURCES AND DATES]. Match the confidence to the evidence: measured facts stated plainly, survey results attributed to their sample ("of 2,000 US adults surveyed by X in 2025"), projections labeled as projections with their source. Never let a projection or survey masquerade as a measurement.

I am presenting [STATISTIC] to [AUDIENCE]. Make it land without distorting it: a comparison that makes the scale graspable ([PER DAY / PER PERSON / VERSUS SOMETHING FAMILIAR]), the one piece of context the audience needs to interpret it correctly, and the honest headline phrasing. Then flag the tempting-but-misleading framings I should avoid even though they sound better.

I am charting this data: [DESCRIBE DATA AND PLANNED CHART]. Check the visualization for honesty: does the axis start at zero or exaggerate the change, is the chart type right for this data relationship, are the intervals consistent, and does the visual impression match the numerical reality? Suggest the honest version, and note what a hostile reader would call out in the current plan.

Build the source documentation for my [REPORT / ARTICLE / DECK]: for every statistic used, list the figure, the original source with full citation and link, measurement date, and one-line methodology note. Order them as they appear. This is my defense file: when someone challenges a number, I answer in thirty seconds instead of scrambling.

Assemble my argument for [POSITION / DECISION] from the verified data we gathered: the three strongest numbers with their sources, the logical chain connecting them to the conclusion, the honest acknowledgment of what the data does not show, and the counter-statistic an opponent would cite with my grounded response to it. An argument that pre-handles the counter-data wins the room.