Tips for Better Results | Arithlab AI Documentation

Getting Better Insights from Your Data

Follow these tips to get the most accurate and useful insights from Arithlab.

Preparing Your Data

Before uploading:

• Use clear, descriptive column headers (e.g., "Sales Revenue" instead of "Rev")

• Make sure dates are in a consistent format

• Remove any test data or incomplete rows

• Include all relevant columns for your analysis

File organization:

• Keep related data in the same file when possible

• Use meaningful file names that describe the content

• Update data regularly for the most current insights

Asking Better Questions

Be specific about what you want:

• Instead of "show me sales" ask "show me monthly sales by product category"

• Include time periods: "last 6 months", "this quarter", "year over year"

• Specify metrics: "revenue", "units sold", "profit margin"

Use follow-up questions:

• Start broad: "What are my top products?"

• Then drill down: "Show me the top products by month"

• Get specific: "Which months had the highest sales for Product A?"

Making the Most of AI Cleaning

When to use different cleaning levels:

• Conservative: For clean data that just needs minor touch-ups

• Moderate: For typical business data with some quality issues

• Aggressive: For very messy data from multiple sources

Review cleaning suggestions:

• Always check what changes the AI recommends

• Understand why certain changes are suggested

• Use the preview to see before/after comparisons

Creating Better Reports

Structure your analysis:

1. Start with high-level overview questions

2. Drill down into specific areas of interest

3. Compare different time periods or segments

4. End with actionable insights and next steps

Export tips:

• Use PDF for executive summaries

• Use PowerPoint for presentation to teams

• Use Excel for detailed data sharing

• Include context and interpretation with your charts

Working with Large Datasets

For files over 10MB:

• Focus on the most important columns first

• Use date ranges to analyze specific time periods

• Break down analysis by categories or segments

• Consider cleaning the data first for better performance

Memory-friendly questions:

• Ask for "top 10" or "top 20" instead of showing everything

• Use time filters: "last 30 days", "this quarter"

• Focus on specific categories or regions

• Use summary metrics instead of detailed breakdowns

Common Mistakes to Avoid

• Asking too many different questions in one query

• Using unclear abbreviations in your questions

• Not cleaning messy data before analysis

• Forgetting to specify time periods

• Not following up to dig deeper into interesting findings