Skip to main content

Data preparation guide

Making your data easy for Dexter to analyze starts with clear, organized, and well-documented files. This guide provides best practices and examples to ensure Dexter can deliver accurate, actionable insights every time.

D
Written by Dominykas Sintra
Updated today

Why data preparation matters

Well-prepared data saves time, reduces errors, and helps Dexter deliver more accurate, relevant results. When you organize your data and clarify your goals, you empower Dexter to focus on what matters most to you—turning raw numbers into meaningful insights.

Step-by-step instructions

Step 1: Define your objective

  • Clarify your goals: Start by writing down what you want to achieve. Are you looking for sales trends, churn analysis, or operational bottlenecks?

  • List your key questions: The more specific your questions, the better Dexter can tailor the analysis.

Example:

  • "Which marketing campaign generated the highest ROI in Q2 2025?"

  • "Can you compare employee turnover rates before and after the new training program?"

Step 2: Gather and organize data

  • Collect all relevant files: Include spreadsheets, reports, exports from software tools, and supporting documents.

  • Ensure completeness: Check that you have all months, categories, or data points needed for your questions.

  • Remove clutter: Delete irrelevant tabs, columns, or outdated data that could confuse the analysis.

Step 3: Clean and Structure your data

  • Standardize formats: Use consistent date, currency, and number formats (e.g., always use YYYY-MM-DD for dates).

  • Label clearly: Use descriptive column headers and avoid abbreviations unless you provide a key.

  • Handle missing data: Fill in gaps where possible, or clearly mark missing values (e.g., with 'N/A').

  • Eliminate duplicates: Scan for and remove duplicate rows or records.

Example:

"The 'Dept' column uses HR for Human Resources, IT for Information Technology—see the key below the table."

Step 4: Provide context and documentation

  • Describe your dataset: Add a short summary explaining what the data covers, time periods, and any special calculations.

  • Highlight important sections: Use comments or notes to direct Dexter to specific tabs, columns, or metrics.

  • Explain abbreviations and formulas: Include a legend or documentation for any shorthand or computed fields.

Example:

"This file contains website traffic data from Jan–Dec 2025. The 'Status' column uses A = Active, I = Inactive. The 'Session Length' column is measured in minutes."

Step 5: Attach all supporting files

  • Upload everything to Brain AI: Make sure all relevant files are available in your workspace before requesting analysis.

  • Include screenshots or past reports: If you have visualizations, charts, or previous analyses, attach these as references.

Step 6: Craft a clear, direct prompt for Dexter

  • Be specific about your needs: Instead of "Analyze my data," say, "Identify which months in 2025 had the highest website conversion rates and explain possible reasons."

  • List any sub-questions: If you need breakdowns by channel, region, or time, include those in your prompt.

Prompt template:

## Questions 
- What are the top conversion months in 2025?
- Are there any patterns by marketing channel?
# Data Context
- File includes monthly website conversions by channel for 2025. Channels are coded as SEO, PPC, Social, etc. All values in USD.
## Output Preferences
- Bullet points, concise explanations, and a summary table if possible.

Did this answer your question?