Data AnalysisLlama
Llama Prompts for Data Analysis
Llama's open-source nature makes it suitable for sensitive data analysis tasks where data cannot be sent to external APIs. Running analysis on-premises with Llama provides the AI capability of cloud models without data leaving your infrastructure.
8 copy-paste ready prompts — optimised for Llama.
8 Llama Data Analysis Prompts
Data Summary Narrative
Analyse this dataset and write a 300-word narrative summary for a non-technical [executive/stakeholder] audience. Data: [paste data or description]. Cover: the key trend, any anomalies or outliers, comparison to [benchmark/previous period], and 2-3 actionable recommendations. Avoid technical jargon.
SQL Query
Write a [database: PostgreSQL/MySQL/BigQuery] SQL query to [describe what you need]. Tables available: [table names and brief descriptions]. Relevant columns: [column names and types]. Requirements: [any specific conditions, joins, groupings]. Include comments explaining each step and an index recommendation if applicable.
Python Data Analysis
Write a Python script using pandas to analyse this dataset: [describe data structure]. Tasks: 1. Load from [CSV/JSON/database]. 2. Clean: handle nulls, remove duplicates, fix types. 3. Compute: [specific metrics or aggregations]. 4. Output: [format: print/CSV/chart]. Include comments and print intermediate results for debugging.
A/B Test Analysis
Analyse the results of this A/B test: [paste results including sample sizes, conversion rates, and confidence level]. Answer: 1. Is the result statistically significant? (show calculation). 2. What is the practical significance (effect size)? 3. Which variant should we ship and why? 4. Any caveats or confounding factors to note?
Dashboard KPI Brief
Design a KPI dashboard brief for a [type of team: marketing/product/sales/operations] team. Business goal: [goal]. Include: 5-7 key metrics with definitions, how each is calculated, data source, target/benchmark, update frequency, and the 1-2 most important metrics that should be prominently displayed.
Data Visualisation Recommendations
Recommend the best chart types to visualise the following data relationships: [describe your data and what you want to show]. For each recommendation: chart type, reason for choosing it, key design considerations (axes, colours, labels), and any pitfalls to avoid. Data: [paste or describe data].
Cohort Analysis Explanation
Analyse this cohort retention data and explain the key findings in plain English: [paste cohort table]. Identify: the strongest and weakest performing cohorts, whether retention is improving or declining over time, the period with the steepest drop-off, and 2 hypotheses for why certain cohorts perform better.
Forecast Summary
Given this historical data: [paste data], generate a 12-month forecast narrative. Include: the forecasting method used and why, key assumptions, the central forecast with a high/low range, factors that could cause the forecast to miss, and how to monitor performance against forecast monthly.
Tips for using Llama for Data Analysis
- For privacy-sensitive data analysis, self-hosted Llama processes data entirely on your infrastructure
- Use structured output (JSON) prompting — Llama follows output schemas reliably with explicit format instructions
- Llama R1-distilled variants are strong for analytical reasoning tasks that require step-by-step derivations
Get a prompt tailored to your exact data analysis task
PromptIt asks a few questions and builds a Llama-optimised prompt for your specific use case — free.
✦ Try PromptIt Free