Add prompt template library for data science workflows

[Copilot]

Creates reusable, model-labelled prompt templates for common data science tasks: exploratory data analysis, model selection, and experiment design.

Changes

• New directory structure: templates/data-science/ for task-specific prompt templates

• EDA template (eda-analysis.md): Production-ready system prompt for exploratory data analysis

  • Instructs LLM to systematically investigate missing values, outliers, and correlations
  • Cites specific pandas functions for each analysis step (e.g., df.isnull().sum(), df.corr(), df.describe())
  • Includes model recommendations (GPT-4, Claude 3 Opus/Sonnet), limitations, and ideal use contexts

• Placeholder templates: model-selection.md and experiment-design.md with metadata structure and future development notes

Usage Example

# Use with LLM API
prompt = """
You are an advanced Data Scientist specializing in EDA.
Analyze customer_data.csv focusing on missing values using df.isnull().sum(),
outliers via IQR method with df.quantile(0.25/0.75), 
and correlations using df.corr().
"""

Each template specifies recommended models, expected limitations (e.g., small samples n<30, high dimensionality >50 features), and contexts where the template should/shouldn't be applied.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Reproducible Prompt Templates for Data Science Workflows</issue_title>
<issue_description>We are curating high-quality, task-specific prompt templates for data science use cases such as exploratory data analysis, feature engineering, model selection, experiment design, and report generation. Templates must be model-labelled, tested, and include notes on expected limitations and ideal contexts of use.</issue_description>

<agent_instructions>et's create a library of prompt templates.

Create a directory templates/data-science/.
Add three file placeholders: eda-analysis.md, model-selection.md, and experiment-design.md.
In eda-analysis.md, write a "System Prompt" designed for an advanced Data Scientist role. It should instruct the model to look for outliers, missing values, and correlation patterns, and require it to cite specific pandas functions to investigate them.</agent_instructions>

Comments on the Issue (you are @copilot in this section)

• Fixes Reproducible Prompt Templates for Data Science Workflows #15

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.