> ## Documentation Index
> Fetch the complete documentation index at: https://docs.isotopes.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Asking questions

# Asking Effective Questions

The quality of your questions directly impacts the quality of aidnn's responses. Learn how to ask questions that get you the insights you need.

## The Anatomy of a Good Question

### Three Key Components

**1. What You Want (The Goal)**
Be clear about what you're trying to accomplish.

❌ "Look at this data"
✅ "Calculate total revenue"

**2. Which Data (The Source)**
Specify what data to use.

❌ "Show me sales"
✅ "Show me sales from my QuickBooks connection"

**3. How to Present It (The Format)**
Describe the output you want.

❌ "Show sales data"
✅ "Show monthly sales totals as a line chart"

### Complete Question Example

> "Calculate total revenue by product category from my 2024 sales file and show it as a bar chart sorted from highest to lowest"

**Breaks down to:**

* **What**: Calculate total revenue by product category
* **Which**: 2024 sales file
* **How**: Bar chart, sorted highest to lowest

## Question Types

### Exploratory Questions

**Purpose**: Understand your data

**Good Examples:**

> "What does this dataset contain?"

> "Show me summary statistics"

> "What's the date range and how many records?"

> "Are there any data quality issues?"

**When to Use:**

* First time working with data
* Unfamiliar dataset
* Before detailed analysis

### Analytical Questions

**Purpose**: Find specific insights

**Good Examples:**

> "What are the top 10 products by revenue?"

> "How does Q4 compare to Q3?"

> "What's the correlation between price and sales volume?"

> "Are there any seasonal patterns in the data?"

**When to Use:**

* You have a specific question
* Looking for patterns or trends
* Making comparisons
* Testing hypotheses

### Action Questions

**Purpose**: Transform or prepare data

**Good Examples:**

> "Remove rows with missing revenue values"

> "Create a new column calculating profit margin"

> "Join this data with my customer file"

> "Export the top customers to CSV"

**When to Use:**

* Cleaning data
* Creating new calculations
* Combining datasets
* Preparing reports

## Being Specific

### Time Periods

**❌ Vague:**

* "Recent sales"
* "Last quarter"
* "This year"

**✅ Specific:**

* "Sales from January 1 to March 31, 2024"
* "Q4 2023 (October through December)"
* "Calendar year 2024"

**Pro Tip:** When you say "this year," aidnn interprets based on current date. Be explicit if you mean something else.

### Metrics and Calculations

**❌ Vague:**

* "Average performance"
* "Total sales"
* "Best customers"

**✅ Specific:**

* "Average order value (revenue / number of orders)"
* "Sum of all revenue in the revenue column"
* "Top 20 customers ranked by lifetime value"

### Filtering and Criteria

**❌ Vague:**

* "Show good customers"
* "Recent orders"
* "Big transactions"

**✅ Specific:**

* "Show customers with lifetime value over \$10,000"
* "Orders from the last 30 days"
* "Transactions where amount is greater than \$1,000"

## Common Pitfalls

### Pitfall 1: Too Vague

**❌ Problem:**

> "Analyze this"

**Why It's Bad:**
aidnn doesn't know what kind of analysis you want.

**✅ Better:**

> "Show me summary statistics and check for missing values"

### Pitfall 2: Multiple Questions at Once

**❌ Problem:**

> "Show me revenue trends, top products, customer segments, and regional performance"

**Why It's Bad:**
Too much at once makes it hard to get focused results.

**✅ Better:**
Ask one at a time:

1. "Show me revenue trends by month"
2. "What are the top 10 products?"
3. Then build from there

### Pitfall 3: Assuming Context

**❌ Problem:**

> "Compare them"

**Why It's Bad:**
aidnn doesn't know what "them" refers to.

**✅ Better:**

> "Compare the monthly totals we just calculated to last year's monthly totals"

### Pitfall 4: Technical Jargon Without Explanation

**❌ Problem:**

> "Run a two-sample t-test on cohort performance"

**Why It's Bad:**
Might work, but simpler language often works better.

**✅ Better:**

> "Test if there's a statistically significant difference between group A and group B performance"

### Pitfall 5: No Output Format Specified

**❌ Problem:**

> "Calculate monthly revenue"

**Not Bad, but Could Be Better:**

**✅ Better:**

> "Calculate monthly revenue and show it as a line chart"

Or:

> "Calculate monthly revenue and display as a table"

## Progressive Questioning

### Start Broad, Then Narrow

**Step 1 - Overview:**

> "Show me a summary of this sales data"

**Step 2 - Identify Interest:**

> "What are the top products?"

**Step 3 - Dig Deeper:**

> "Show me weekly sales trend for the #1 product"

**Step 4 - Specific Analysis:**

> "Why did sales spike in week 3?"

### Build on Previous Answers

**Use Context:**
Instead of:

> "Calculate average order value for transactions between $0 and $100 in Q4 2023"

**Build Step by Step:**

1. "Filter to Q4 2023 transactions"
2. "Now filter to amounts between $0 and $100"
3. "Calculate average order value for these orders"

aidnn remembers each step and applies them cumulatively.

## Question Patterns That Work

### Pattern 1: Describe Then Request

**Structure:**

1. Describe what you have
2. State what you want

**Example:**

> "I have sales data with customer, product, date, and revenue. Show me total revenue by product for Q4 2023."

### Pattern 2: Goal Then Format

**Structure:**

1. State your goal
2. Specify output format

**Example:**

> "I need to identify our top customers. Show me the top 20 customers by total revenue as a table with customer name, total spent, and number of orders."

### Pattern 3: Filter Then Analyze

**Structure:**

1. Filter the data
2. Analyze the subset

**Example:**

> "Filter to transactions in California from 2024. Then calculate monthly revenue totals and create a trend chart."

### Pattern 4: Compare Then Contrast

**Structure:**

1. Define what to compare
2. Specify comparison method

**Example:**

> "Compare Q1 2024 revenue to Q1 2023 revenue. Show both totals and calculate the percentage change."

## Advanced Techniques

### Conditional Logic

**Use "if/then" structure:**

> "If revenue is greater than \$10,000, label it as 'high-value'. Otherwise, label as 'standard'."

### Multiple Criteria

**Combine conditions with AND/OR:**

> "Show me customers where total spend is over \$5,000 AND last purchase was within 30 days"

### Grouped Analysis

**Specify dimensions to group by:**

> "Calculate average order value grouped by customer segment and month"

### Time-Based Comparisons

**Use relative time periods:**

> "Compare this month to same month last year"

> "Show year-over-year growth for each quarter"

## When aidnn Doesn't Understand

### Rephrase More Simply

**If aidnn seems confused:**

**Original:**

> "Elucidate the longitudinal revenue trajectory"

**Simpler:**

> "Show me revenue over time"

### Break It Down

**Instead of one complex question:**

> "Perform cohort analysis with retention curves and revenue per cohort showing month-over-month changes"

**Break into steps:**

1. "Create customer cohorts by signup month"
2. "Calculate retention rate for each cohort"
3. "Show retention rates over time"
4. "Add revenue per cohort"

### Provide Examples

**If aidnn isn't getting it:**

> "I want to see each product with its category. For example: 'Widget A, Electronics' or 'Gadget B, Home Goods'"

### Ask for Clarification

**Let aidnn ask back:**

> "Show me customer performance"

aidnn might ask:

> "What metric would you like to see for customer performance? Options include:
>
> * Total revenue per customer
> * Number of orders per customer
> * Average order value
> * Customer lifetime value"

## Quick Checklist

Before sending your question, verify:

* [ ] Is my goal clear?
* [ ] Did I specify which data?
* [ ] Did I mention the time period (if relevant)?
* [ ] Did I describe the output format?
* [ ] Am I asking one thing at a time?
* [ ] Did I provide enough context?
* [ ] Is my language clear and specific?

## Examples: Before and After

### Example 1: Sales Question

**❌ Before:**

> "Sales?"

**✅ After:**

> "Show me total sales by month for 2024 from my QuickBooks data"

### Example 2: Customer Analysis

**❌ Before:**

> "Who are the good customers"

**✅ After:**

> "Show me the top 25 customers ranked by total lifetime revenue, including customer name and total spent"

### Example 3: Trend Analysis

**❌ Before:**

> "Chart it"

**✅ After:**

> "Create a line chart of the monthly revenue totals we just calculated"

### Example 4: Comparison

**❌ Before:**

> "How does this year compare"

**✅ After:**

> "Compare 2024 monthly revenue to 2023 monthly revenue and calculate the percentage change for each month"

## Practice Exercise

**Try turning these vague questions into specific ones:**

**Vague:** "Show me the data"
**Your Version:** \_\_\_\_\_?

**Vague:** "What's the trend"
**Your Version:** \_\_\_\_\_?

**Vague:** "Best products"
**Your Version:** \_\_\_\_\_?

**Suggested Answers:**

1. "Show me the first 50 rows of my sales data with all columns"
2. "Show me monthly revenue trend for the past 12 months as a line chart"
3. "Show me the top 10 products by total revenue for Q4 2023"

## Next Steps

* **[Working with Large Datasets](./large-datasets)** - Handle big data efficiently
* **[Use Cases](../use-cases/sales-analysis)** - See examples in action
* **[Conversations Guide](../using-aidnn/conversations)** - Master interactions

Questions? Check our **[FAQ](../help/faq)** or reach us at [support@isotopes.ai](mailto:support@isotopes.ai).