This project provides a research analytics environment with integrated AI support, memory systems, and reproducible analysis workflows. It combines data analysis capabilities with advanced AI assistance for social science research. This project uses public data on Ukrainian publishing trends sourced from the Books Chamber of Ukraine and prepared for analysis by github.com/rg-fides/books-of-ukraine.

• Project Memory System: Separate memory tracks for human decisions and AI technical notes, with summary and detailed views
• Context Management: Dynamic AI context loading for specific analytical tasks (i.e. flexible system prompt)
• Copilot Integration: GitHub Copilot optimized for research analytics workflows (e.g. tasks, extensions)

• Reproducible Workflows: Scripted analysis pipelines with version control
• Template System: Pre-built templates for common research tasks
• Documentation Integration: Seamless integration of analysis and documentation
• Multi-Format Output: Reports in various formats (HTML, PDF, etc.)

1. Setup: Install VS Code and activate Copilot
2. Fork Repository: github.com/andriy.koval/aim-2025-sandbox
3. Environment: Work with AI agent to install dependencies
4. Explore project: Request project briefing from Copilot
5. Analysis: Open analysis/eda-1/eda-1.R and start exploring the data with AI assistancenced Research Analytics

5. Analysis: Open ./analysis/eda-1.R and start exploring the data with AI assistance.

Database Access: Two-stage database system with prepared data from books-of-ukraine project:

• connect_books_db("main") - Final analytical database (analysis-ready)
• connect_books_db("stage_2") - Books + admin + custom data

Quick Start: Run Rscript scripts/test-database-connection.R to verify database access

Analysis Templates: Ready-to-use templates in analysis/ directory
AI Memory System: Use ai/memory-hub.md to manage research insights
Context Management: Dynamic AI assistance via the ./ai/ system
Documentation: Comprehensive guides in guides/ directory

Some development tasks in .vscode/tasks.json invoke PowerShell scripts. To allow these to run on machines with restricted PowerShell execution policy, the tasks include the flags -ExecutionPolicy Bypass -NoProfile when invoking PowerShell. This does not change the system's execution policy; it only instructs the PowerShell process started by VS Code to bypass the policy for that run.

If you prefer to change your user execution policy permanently instead, run PowerShell as Administrator and set:

Set-ExecutionPolicy -Scope CurrentUser -ExecutionPolicy RemoteSigned

Only do that if you understand the security implications. The current task approach is minimally invasive and safe for development workflows.