AI Prompt Library

You are my data-analysis partner. Here is my data (paste a table or describe the columns and samples):
(data)

What I want to know: (e.g., "which marketing channel has the highest ROI, and how should I split next month's budget")

Please:
1. First list how you'd break this question down (3-5 analysis angles).
2. Point out what the existing data can and can't answer.
3. Give your interpretation and one concrete, actionable recommendation.
4. Remind me of this conclusion's premises and risks.

You are a senior Excel analyst. I want to do the following calculation in (cell location): (describe in plain language what to compute, e.g., sum column B amounts by column A department). Give me a paste-ready formula, explain each function's role in one line, and if there's a more robust way (e.g., SUMIFS instead of SUMIF) provide that too. My columns are: (paste the first two rows' headers and sample data).

You are a data-analysis consultant. Below are my data columns and a small sample: (paste column names and 3 to 5 rows of data). My analysis goal is (the business question to answer). Suggest the best pivot-table configuration: which columns to put in rows, columns, values, and filters, and list step-by-step how to do it in Excel or Google Sheets, and finally point out 3 insight directions this pivot might reveal.

You are a database engineer. My table schema is as follows: (paste table names and each column's name and type). I want to query: (describe in plain language, e.g., find each customer's total spend over the past 30 days, sorted high to low). Write standard SQL (state the database type, e.g., MySQL or PostgreSQL), with comments, and warn me of possible performance traps and suggested indexes.

You are a data-cleaning expert. Below is a sample of my raw data: (paste 10 rows with problems). Please: 1) list the data-quality issues you find (missing, duplicate, inconsistent format, outliers). 2) for each issue, give a concrete handling method and criterion. 3) provide cleaning steps executable in Excel or Python (pandas) so the data is ready for analysis.

You are a market-research analyst. Below are the statistical results of (survey topic): (paste each question's options with percentages or counts). Produce a summary: three sentences of overall highlights, findings for each key question, differences across groups (e.g., age or gender), and 3 actionable recommendations. Objective tone; cite the source questions for numbers, don't over-infer.

You are a qualitative researcher. Below is a list of open-ended survey responses: (paste multiple text responses). Group them into 5 to 8 thematic categories, each with a clear name and definition, tally each category's approximate share, and attach 2 representative original quotes each. Finally, point out the one signal most worth attention and why.

You are a business data analyst. Below is recent data for (metric name, e.g., monthly revenue): (paste each period's value). Interpret the trend: overall direction, growth or decline magnitude, possible turning points and hypothesized underlying causes. Then write it as a narrative for a report so a non-data manager gets it, and finally raise 2 questions needing further validation.

You are a growth analyst. I ran an A/B test: control group sample size (number), conversions (number); treatment group sample size (number), conversions (number). The test goal is (the metric to improve). Compute both groups' conversion rates and the relative lift, judge whether the difference is statistically significant (explain the method and p-value concept used), and give a clear "should we roll out to everyone" recommendation with reasons.

You are a Python data analyst. I have a CSV with columns: (paste the column names). I want to (describe the analysis or chart need). Write complete, runnable code using pandas and matplotlib, including reading the file, handling missing values, computing, and plotting, with Chinese comments per section, and finally explain how to interpret the output.

You are an operations-data consultant. My business is (brief industry and business model), and the main goal is (e.g., raising repurchase rate). Design a KPI dashboard: list 6 to 10 key metrics, each with its definition, calculation, tracking frequency, and healthy range, grouped into "leading indicators" and "lagging indicators," and finally suggest the dashboard's layout priority.

You are a data-visualization expert. Below is the data I want to present: (paste the data or describe its type and comparison purpose). My audience is (audience), and I want to emphasize (the key point to convey). Suggest the most suitable chart type and why, explain which axes to use and how to color and sort without misleading, and warn me of common visualization mistakes, so the chart is instantly understandable.

You are a strategy analyst. Help me build a competitor comparison analysis. We are (our brand), the competitors are (list 2 to 4), and the comparison dimensions include (e.g., price, features, audience, channels, reviews). Organize into a table, comparing item by item objectively, marking our strengths and weaknesses, and finally summarize 3 differentiation opportunities to enter. If you need me to add data, list what first.

You are a data-audit specialist. Below is a set of (metric name) data: (paste the value series or dated data). Find possible outliers or implausible spots, explain your basis for judging them abnormal (e.g., outliers, sudden spikes/drops, inconsistency with the trend), guess possible causes (input error, seasonal factors, real events), and suggest whether each anomaly should be kept, corrected, or verified further.

You are a financial analyst. Below is (a company or project)'s financial data: (paste revenue, cost, gross profit, net profit, etc., possibly multi-period). Compute key ratios (gross margin, net margin, growth rate, etc.), point out the financial health's highlights and warning signs, compare changes across periods and guess causes, and finally summarize the overall health in plain language a non-finance person understands, plus the one thing to watch.

You are a senior analysis-report writer. I'm writing a data-analysis report on (analysis topic) for (audience); the core finding is (one-sentence conclusion), and my data includes (brief data description). First plan the report's chapter structure (from summary to recommendations), explain what content and charts go in each section, and demonstrate how to write the opening executive summary, achieving conclusion-first with data support.

You are a data-forecasting analyst. Below is historical data for (metric name): (paste each period's value). Forecast the next (periods) periods based on the trend, explaining the estimation logic and assumptions used, and provide numbers and premises for optimistic, baseline, and conservative scenarios. Finally, note this forecast's biggest source of uncertainty and which signal to track to revise the judgment.

Below is some data. Help me spot the key trends, outliers, and possible reasons behind them, and finally give me 3 "questions worth digging into further."

Data: "___"

In a spreadsheet I want to achieve "___ (need)", with columns "___". Write the corresponding formula (Excel/Google Sheets) and explain what each part does so I can modify it myself later.

Below are my data results for this month. Write them into a "plain-language summary" for a manager: what happened, why, and what I recommend doing next. Don't just dump numbers.

Data: "___"

In a spreadsheet I want to achieve: (describe the need, e.g., "a monthly sum by dates in column B"). Give me a working Excel and Google Sheets formula, and explain the role of each part.

You are a spreadsheet ace. Below is a formula I don't understand — help me break it down.

Answer like this:
1. [What this formula does] State in one plain sentence what it ultimately computes.
2. [Section-by-section breakdown] Split the formula into a few parts and explain each part's role.
3. [Any problems] Check for potential errors, failure situations (e.g., division by zero, value not found), or a simpler way to write it.
4. [Rewrite suggestion] If there's a clearer or more robust version, give one and explain the difference.

Software I use: (Excel or Google Sheets)
Formula: (paste the formula)
The context this formula sits in: (briefly what this sheet computes)

Explain in a way even a beginner understands, with minimal jargon.

I have a very messy spreadsheet and want to tidy it into a clean, analyzable format. From a data analyst's view, tell me step by step how to clean it.

Roughly what the data looks like: (describe the columns, e.g., date/product/quantity/amount, or paste the first few rows)
The question I ultimately want this data to answer: (e.g., which month sold best, which product had the most returns)
Problems I've found so far: (e.g., inconsistent date formats, blank rows, inconsistent category names)

Please:
1. List the items needing cleaning (as a list, ordered by what to handle first)
2. Give a concrete handling method for each (manual or with a tool)
3. After cleaning, suggest 2-3 angles to slice the data to answer my question
4. Remind me of common data traps to watch for (e.g., double counting, inconsistent units)

Below is my data:
(paste numbers/table)

Find the main trends, anomalies, and noteworthy insights, and summarize 3 key points and possible action recommendations in plain language.

I have this data: (describe the columns and content). I want to convey (the key point). Suggest the most suitable chart type and explain why.

The data below is messily formatted:
(paste)

Tidy it into a unified, clean table format (consistent columns, removing duplicates and extra symbols), and explain what adjustments you made.

You are a database engineer. I'll describe in plain language what data I want to query — write the corresponding SQL query.

Database type: (fill in MySQL / PostgreSQL / BigQuery, etc.)
Relevant tables and columns: (paste table names and main columns, e.g., the orders table has order_id, customer_id, amount, created_at)
What I want to query: (describe in plain language, e.g., "each customer's total spend last month, sorted high to low")

Please:
1. Give me directly-runnable SQL, with short comments at key spots explaining what it does.
2. If my description is ambiguous, point it out and ask me rather than assuming.
3. If this query might be slow, remind me how to optimize (e.g., adding indexes).

I want to compute some result in a spreadsheet but don't know the formula — translate my need into a paste-ready formula.

I use: (Google Sheets or Excel)
What my data looks like: (describe the columns, e.g., column A dates, column B product, column C amount; or paste the first few rows)
The result I want: (describe in plain language, e.g., the total amount per product per month / the count matching a condition)

Please:
1. Give me a paste-ready formula and clearly mark which cell it goes in
2. Explain section by section what each part of the formula does
3. Remind me of common error spots (e.g., data range, absolute references, blank cells, inconsistent formats)
4. If there's a simpler or more robust approach (e.g., a pivot table), suggest that too

Constraint: follow my specified software's syntax; don't get function names wrong.

I want to understand (survey purpose). Design an 8-10 question survey with varied question types (choice, scale, open-ended), and explain what information each question aims to get.

I've uploaded a (describe the data, e.g., monthly sales data by channel over the past year). I'm not great with formulas — please: 1) first summarize in a paragraph what this data is about, how many records, and which columns it covers; 2) answer my questions (list your questions, e.g., which channel grew fastest? which month was abnormal?); 3) produce suitable charts; 4) proactively tell me 3 insights or anomalies "I may not have noticed but are worth attention." Explain in plain language, don't just dump numbers.

Below is a chart's data/description: (paste). Help me interpret what it's saying and the most important findings, and point out possible interpretation traps.

You are a data analyst skilled at putting numbers into human terms. I'll paste some data (a table or description) — help me interpret it.

Data: (paste your data, e.g., monthly revenue over the past half year, conversion rate by channel)
What I want to know: (e.g., where it's growing, where there's a problem, what to focus on)

Please:
1. In three to five plain-language points, make clear what this data is saying.
2. Point out the most notable anomaly or trend and guess the possible cause (mark it as a guess).
3. Give me 2-3 directions for further questions or action.
Don't pile on jargon — explain as if to a colleague who doesn't know data. For important conclusions, remind me to go back and verify against the raw data.

Below is inconsistently-formatted data: (paste). Tidy it into a unified format (consistent columns, deduplicated, filled in), and explain what handling you did.

I ran a survey and got a pile of open-ended text responses, and I want to find what people are actually saying rather than reading each one until my eyes glaze over.

The survey's topic and purpose: ()
The question I asked: (paste the question)
Responses:
(paste all open-ended responses, one per line is fine)

Please:
1. Summarize these responses into a few main themes, each with an understandable name
2. Mark roughly what share each theme accounts for (use "most / some / few" — don't fake precise numbers)
3. Pick 1-2 most representative original responses as examples per theme
4. Point out 1-2 "nobody said but worth noting" signals, and give me one summary line I can report to the team.

This is our company's income statement or income-expense data for the past (months) months (paste the data). I'm not from a finance background — please: 1) summarize the period's operating status in three plain sentences (revenue trend, gross-margin change, ultimately profit or loss); 2) check costs and expenses item by item, find items growing faster than revenue, and mark the three to prioritize; 3) run a "if this trend continues, what happens in three months" projection; 4) give me three concrete actions to execute next month. Translate all jargon into plain language when explaining.

Below is my data: (paste). Find the 3 most important trends or findings, explain them in plain language, and give a possible action recommendation for each.

You are a household finance analyst. Below is my spending record for this month (de-identified): (paste amounts + brief running records or statement details). Please: 1) auto-categorize and list each category's amount and proportion; 2) point out the top three spending patterns I may not have realized; 3) find all subscriptions and recurring expenses and flag those with questionable usage frequency; 4) give me "one" concrete, doable improvement goal for next month. No investment advice.

I want to manage "(what data)". Design a spreadsheet column structure, what goes in each column, and suggestions to make it easy to tally.

I'm considering (option A/B/C), weighing factors like (cost, benefit, risk...). Make a comparison table and give a recommendation with reasons based on my situation.

Below is my data: (paste). Check for outliers, missing values, duplicates, or inconsistent formats, and point out spots that may affect the analysis.

Below are several things I want to compare: (paste). Organize them into a clear comparison table, using the most critical comparison dimensions as columns.

My business is (briefly), current stage (). List the 5 key metrics (KPIs) I should most track at this stage, and explain why each matters and how to read it.

Below is my research data/statistical results: (paste). Explain in plain language what these numbers mean, the most important findings, and the traps to watch when interpreting.

Below is data for my ad channels (paste ROAS, spend, orders, CAC, etc.). Judge: which channel is profitable, which is losing money, where to shift budget, and remind me whether I've overlooked costs like returns, shipping, and payment-processing fees. Finally, give me 3 concrete adjustments for this week.

These are (company name)'s latest quarter earnings highlights (paste: revenue, gross margin, operating income, EPS, YoY growth, etc.). I'm an investing beginner — interpret in plain language: 1) is this report overall good or bad? The key changes vs. last quarter and the same period last year; 2) what the changes in gross margin and operating margin reveal; 3) which numbers look good but actually warrant caution (e.g., non-operating income contribution); 4) which three metrics I should track next. Finally, remind me this is not investment advice, and I should read the original filing on the Market Observation Post System before major decisions.

I've finished an analysis; the conclusion is: (paste your technical conclusion, including statistical figures, model metrics, jargon). The audience is (who: a department manager who doesn't read code / a client / a coaching staff), and what they care most about is (the decision question: whether to invest money, whether to change approach, how big the risk is). Rewrite it in three lengths: 1) a one-sentence version (no jargon at all); 2) a three-point version (each with a concrete number and what it actually means); 3) a five-minute verbal-report version (with an analogy they're familiar with, plus the three questions I should expect and my answers). For any inference my data doesn't support, mark it directly as "the data can't support this point" — don't cover for me. In Traditional Chinese with Taiwan usage.

I want to draw a conclusion from this data: (describe your data, e.g., 40 games in a season, 260 orders over 3 months for a product, an A/B test with 500 users each). The conclusion I want to claim is: (paste your conclusion). Please check: 1) whether this sample size can support a conclusion of this strength, explained in plain language; 2) which common statistical traps might make me see a false signal (regression to the mean, survivorship bias, multiple comparisons...), matched one by one against my situation; 3) if the sample is insufficient, what strength of claim I could make instead to be honest; 4) what more data I'd need to fully support this conclusion. In Traditional Chinese with Taiwan usage, without obscure statistical jargon.

I have a pile of scattered operations records I want to organize into an analyzable structured data table. Be my data-engineering consultant.

[My industry] (e.g., a fish farm / a secondhand clothing shop / a renovation crew)
[How I record now] (e.g., written in a notebook daily, reported in a LINE group, Excel but everyone's format differs)
[Questions I want to answer] (list 3-5, e.g., which month had the highest loss? which worker has the highest add-on rate?)

Output:
1. **Suggested table structure**: column name | data type | required or not | filling instructions | example value. Design it at a complexity "the people on-site are willing to fill in" — don't design twenty columns.
2. **Which columns must exist to answer my questions**: explain one to one
3. **How to convert existing records**: how should I organize past notebooks or LINE messages into this table? Give practical steps
4. **Minimal daily-entry flow**: design a filling action completable within 60 seconds
5. **Analysis possible after three months**: list three charts you could produce then and the questions they'd answer

Be pragmatic. I have no engineer, only Excel or Google Sheets.

You are a data engineer producing a batch of test data containing no real personal data for the team. Help me write a spec sheet.

Table purpose: (e.g., an e-commerce orders table)
Main columns: (e.g., order ID, member ID, name, phone, address, product, amount, order time, payment method)
Data volume needed: (e.g., 500,000 rows)
Purpose: (e.g., performance load-testing + front-end demo)
Industry and regulations: (e.g., e-commerce, must comply with the Personal Data Protection Act)

Output:
1. A column-classification table: which are direct identifiers, indirect identifiers, or non-personal-data, with a suggestion to mask, pseudonymize, or fully synthesize each
2. Logical constraints that must hold between columns (e.g., ship time must be later than order time, postal code must match the city)
3. A list of boundary cases to deliberately create (null values, overly long strings, extreme amounts, cross-year/cross-month, duplicate orders)
4. Post-generation quality acceptance criteria: which statistical metrics to compare and the tolerance allowed
5. The privacy-testing method and pass criteria, and what evidence to keep for auditing

Output the first two items as tables, the rest as bullets.

I run operations for (business type, e.g., e-commerce/SaaS/physical store), and the data I currently have includes: (orders/traffic/members/ads and other data sources). Plan a monthly-report-level KPI dashboard: 1) using a three-tier "north-star metric — leading indicators — lagging indicators" structure, pick 8-12 metrics, explaining why each matters and how to gauge its healthy range; 2) point out which metrics are easily gamed or misread; 3) suggest the dashboard's layout zones and chart types (line for trends, stacked for composition, etc.).

Below is a large amount of customer feedback (reviews/surveys). Please: 1) categorize into positive and negative themes; 2) find the three most-mentioned strengths and three pain points; 3) quote one or two representative originals each; 4) give actionable improvement suggestions and priority. If the data contains personal data, I've de-identified it.

Feedback:
"paste the feedback"

Below is a dataset. Please: 1) find the 3 most notable findings; 2) explain in plain language what each means and why it matters; 3) suggest a suitable chart to present it; 4) give a headline-worthy key conclusion. For complex statistics, remind me to verify myself.

Data:
"paste the data or description"

I want to make a survey — please: 1) design clear, non-leading questions based on my purpose; 2) mix choice, scale, and open-ended questions; 3) order them logically; 4) remind me to avoid bias-inducing phrasings.

Survey purpose: "what I want to understand"
Audience: "who fills it in"

You are an energy-management consultant. I'll provide electricity-usage data — help me find possible waste points.

[My situation]
- Facility type: (office / store / factory / restaurant)
- Number of facilities: (1 / multiple; if multiple, say how many)
- Business hours: (e.g., Mon-Fri 9:00-18:00)
- Data range: (e.g., monthly usage over the past 12 months / hourly time-of-use data over the past 3 months)
- Electricity unit price: (about NT$X per kWh)
- Main power-consuming equipment: (e.g., AC, freezers, lighting, servers)

[Data]
(paste the meter or bill data here)

[Output in order]
1. **Anomaly list**: list the anomalies you find, each stating "what phenomenon was observed" and "possible cause." Pay special attention to base-load usage outside business hours, seasonally-unexplainable changes, and outliers across multiple sites.
2. **Amount conversion**: estimate each anomaly's "annual potential savings (NT$)," explaining the basis and assumptions.
3. **Priority**: rank by "savings ÷ improvement difficulty" as a table with columns: rank / anomaly / annual savings / estimated improvement cost / payback months.
4. **Verification method**: for the top three, propose a "free way to first confirm whether the judgment is right" check each.
5. **Data gaps**: clearly point out which judgments have lower confidence due to insufficient data and what data is needed to confirm.

Answer in Traditional Chinese with Taiwan usage; be sure to mark numbers as estimates with their assumptions — don't give me precise numbers with no basis.

You are a data-analysis consultant. I want to build a BI dashboard — first help me clarify the requirements before giving a spec.

[Basic situation]
- Company type and size:
- Who this dashboard is for: (e.g., the boss / marketing lead / store manager)
- How they currently ask about numbers: (list the most-asked questions)
- Where the existing data is: (e.g., a MySQL orders table, GA4, Excel)
- Update-frequency need: (real-time / daily / weekly)

[Output in order]
1. **Clarify questions**: first list 5 questions you think I haven't made clear but that affect the design. For each, give me 2 to 3 common options to answer quickly.
2. **Metric-definition table**: translate users' colloquial questions into clear metric definitions. Columns include: colloquial phrasing / formal metric name / calculation logic / data source / common misunderstanding. This step matters most — whether "revenue" includes tax, whether returns count, if unclear the rest is all wrong.
3. **Dashboard layout**: suggest which charts to include, what chart type each uses, and why. Order top to bottom by importance, and state what the manager should see within three seconds.
4. **What not to include**: list common charts nobody actually looks at, and explain why to drop them.
5. **Auto-push settings**: suggested send frequency, recipients, and a summary-text template.
6. **Future expansion**: what phase two could add, deliberately left out of v1.

Answer in Traditional Chinese with Taiwan usage; metric definitions should be precise enough to write directly as SQL.