7 Claude Prompts for Data Analysis, Excel and Finance

Write a spreadsheet formula for me.

App and version: [EXCEL 365 / OLDER EXCEL / GOOGLE SHEETS]
What it should do: [PLAIN ENGLISH, E.G. TOTAL EACH REP'S SALES FOR THE CURRENT MONTH, EXCLUDING REFUNDS]
My data layout: [SHEET NAMES, COLUMN LETTERS AND HEADERS, WHICH ROWS HOLD DATA, AND WHICH CELL THE FORMULA GOES IN]
Sample rows: [PASTE 5 TO 10 ROWS, WITH SENSITIVE VALUES REPLACED]

Please:
1. Give the formula ready to paste, using my actual cell references.
2. Explain how it works piece by piece in plain language, so I can adjust it myself later.
3. Handle the edge cases: blank cells, text where numbers are expected, errors, duplicates, dates stored as text, and rows added later. Say how the formula behaves in each.
4. If a simpler or sturdier approach exists, such as a pivot table, a helper column, XLOOKUP instead of nested IFs, or FILTER or QUERY in Sheets, mention it and when to prefer it.
5. Show the expected result for two of my sample rows so I can confirm it works.

Use only functions available in the app and version I named. If my description could mean two different things, ask me before writing the formula.

Analyse this dataset.

<data>
[PASTE THE CSV OR TABLE, OR ATTACH THE FILE]
</data>

What the data is: [SOURCE, WHAT EACH ROW REPRESENTS, TIME PERIOD]
My question: [E.G. WHY DID SALES DROP IN Q3?]
Who the results are for: [E.G. MY MANAGER, A CLIENT, JUST ME]

1. Check the data first: rows, columns, data types, missing values, duplicates, obvious errors, and anything that limits what we can conclude. If a problem is serious, tell me before analysing.
2. Summary statistics for the columns relevant to my question.
3. Trends and patterns over time or across segments, with the numbers behind each.
4. Anomalies and outliers, with possible explanations labelled as hypotheses.
5. A direct answer to my question with your confidence level. Separate what the data shows from what it only suggests, and treat correlation as correlation.
6. Caveats: sample size, missing context, seasonality, and the data that would make us more certain.
7. Two or three charts that would communicate the findings, with chart type, axes and the point each one makes.

If you can run code, use it for the calculations and show your work, so every number is computed rather than estimated. Open with a three-bullet summary written for the audience I named.

Help me design a dashboard that drives decisions rather than just displaying numbers.

The business question or decision it supports: [E.G. WHERE ARE WE LOSING TRIAL USERS, WHICH STORES NEED ATTENTION THIS WEEK]
Who will use it and how often: [ROLES, AND DAILY / WEEKLY / MONTHLY]
Data available: [SOURCES, TABLES, KEY FIELDS]
Tool: [E.G. LOOKER STUDIO, POWER BI, TABLEAU, EXCEL, METABASE]

1. Restate the decisions each user will make with this dashboard. If my question is too vague to design for, ask me to sharpen it.
2. Choose 3 to 7 KPIs. For each: an exact definition and formula, the data source, the comparison that gives it meaning (last period, target or benchmark), and the action someone takes when it moves. Drop any metric that wouldn't change a decision.
3. Separate leading indicators from lagging ones.
4. Layout from top to bottom: headline numbers, then the charts that explain them, then detail tables. Name each chart type and why it fits the data.
5. The filters and drill-downs users will actually need.
6. Data quality risks and definitions people are likely to argue about, such as what counts as an active user, with a suggested definition for each.

Output a top-to-bottom wireframe description followed by a metric dictionary table.

Build a simple financial model for [BUSINESS OR PROJECT] covering [12 MONTHS / 3 YEARS].

<inputs>
Revenue model: [HOW YOU MAKE MONEY, PRICES, PLANS]
Current numbers: [CUSTOMERS, REVENUE, GROWTH RATE, CHURN, IF ANY]
Costs: [FIXED COSTS, COST PER SALE OR CUSTOMER, SALARIES, MARKETING SPEND]
Starting cash and funding: [AMOUNTS]
</inputs>

1. Put every assumption in a separate assumptions table: value, unit, source (my data, my estimate, or your placeholder) and confidence. Mark your placeholders clearly so I replace them.
2. Build the model monthly: customers or units, revenue, direct costs, gross margin, operating costs, profit or loss, and cash balance.
3. Show three scenarios (conservative, base and optimistic) by changing only the two or three assumptions that matter most, and name them.
4. Key outputs: the break-even month, months of cash runway, and the lowest cash point.
5. Sensitivity: which single assumption moves the result the most.

Keep formulas simple and linked to the assumptions table so I can change inputs. If you can create a spreadsheet file, build one with the assumptions on their own tab; otherwise give me tables I can paste into a spreadsheet. Treat this as a planning tool rather than financial advice, and flag anything I should review with an accountant.

Help me review my spending and build a monthly budget. I want practical help, not judgement.

<money>
Monthly take-home income: [AMOUNT, AND WHETHER IT VARIES]
Spending: [PASTE ONE OR TWO MONTHS OF TRANSACTIONS, OR LIST CATEGORIES AND AMOUNTS]
Debts: [BALANCE, INTEREST RATE AND MINIMUM PAYMENT FOR EACH]
Savings and emergency fund: [AMOUNTS]
Goals: [E.G. CLEAR A CREDIT CARD, SAVE A HOUSE DEPOSIT, STOP RUNNING OUT BEFORE PAYDAY]
</money>

1. Sort my spending into essentials, flexible spending, and savings or debt payments, with totals and each one's share of income.
2. Point out patterns: recurring subscriptions, small frequent purchases that add up, and irregular costs I should set aside for monthly (annual bills, gifts, car repairs).
3. Suggest three to five specific changes ranked by money saved against effort, noting which ones might be hard for me, so I can choose.
4. Build a monthly budget table that includes some guilt-free spending money, because budgets with no flexibility tend to get abandoned.
5. If I have debts, compare paying the highest interest rate first with paying the smallest balance first, and what each means in my case.
6. A weekly check-in routine that takes ten minutes or less.

Explain your reasoning in plain language. This is general budgeting help, so tell me when something calls for a qualified adviser or a free debt advice service, especially for tax, investments or serious debt.

Help me research [COMPANY NAME AND TICKER] in a structured way. I'm learning to analyse companies, so this is for understanding rather than a recommendation, and I don't want price predictions.

<materials>
[PASTE SECTIONS OF THE ANNUAL REPORT, AN EARNINGS RELEASE, AN INVESTOR PRESENTATION, OR YOUR NOTES]
</materials>

1. Business model: how the company makes money, its main segments and customers, and what drives revenue. Use the materials I pasted and cite the section.
2. Financials to check: the key metrics for this type of business (for example revenue growth, gross and operating margin, free cash flow, debt, share dilution, segment trends), what each one tells me, and the values from my materials where available. Where data is missing, tell me where in the filings to find it instead of estimating.
3. Competitive position: what protects the business, if anything, and how durable that looks.
4. Risks: company-specific, industry and macro, including the ones the company lists itself.
5. Bull case and bear case: the strongest honest argument for each, and the evidence over the next year or two that would prove each one right.
6. The questions I should answer before forming my own view.

Keep facts from my materials separate from your interpretation, and say where your general knowledge may be out of date. This is not investment advice; remind me to consider a licensed financial adviser before making investment decisions.

Clean this messy data and turn it into a structured table.

<raw_data>
[PASTE THE DATA OR ATTACH THE FILE]
</raw_data>

Target columns and formats: [E.G. NAME, EMAIL IN LOWERCASE, PHONE IN INTERNATIONAL FORMAT, COUNTRY CODE, DATE AS YYYY-MM-DD]
What it's for: [E.G. A CRM IMPORT, ANALYSIS, A MAILING LIST]

Steps:
1. Profile the data first: row count, the formats you see, and the main problems (inconsistent case, mixed date formats, extra spaces, merged fields, missing values).
2. Normalise each field to the target format.
3. Deduplicate: state the rule you used to treat two rows as duplicates (exact match, same email, or similar name plus same company) and keep the most complete version.
4. Flag instead of guessing: when a value is ambiguous (03/04 could be 3 April or 4 March), invalid or missing, leave it blank or unchanged and list it in an issues table with row number, field, original value and reason.
5. Keep a change log of the rules you applied so I can repeat them next time.

Output the clean table as CSV, then the issues table, then the change log with counts: rows in, duplicates removed, rows flagged, rows out. For large files, write and run code so every row is handled the same way.