The Observer

Research and analysis workspace for newspaper discourse and dispute-signal modeling, focused on Observer corpus pipelines and downstream validation/visualization outputs.

Project Overview

This repository contains:

• Corpus collection and filtering pipelines
• Dispute-signal extraction and sentence classification workflows
• Extended actor/spatial/network analysis scripts
• Validation and visualization generators for publication-ready outputs
• Draft manuscript and chapter materials

Repository Layout

• paper2_online_discourse/ — core scripts, research writeups, and generated datasets/figures
• paper2_online_discourse/outputs* — run-specific output directories (raw/collected articles, monthly signals, classified sentences, summaries)
• paper2_online_discourse/outputs_observer_full_corpus/ — full-corpus integrated outputs, tables, and extended analysis artifacts

Quick Start

1. Clone the repository.
2. Create and activate a Python environment.
3. Install required dependencies used by the scripts.
4. Run the target pipeline script from paper2_online_discourse/.

Example script entry points:

• paper2_online_discourse/observer_full_corpus_2015_2025.py
• paper2_online_discourse/pipeline_newspaper_disputes.py
• paper2_online_discourse/observer_historical_pipeline.py
• paper2_online_discourse/generate_observer_visualizations.py

Key Artifacts

• articles_raw.csv, articles_collected.csv
• monthly_dispute_signals.csv
• sentences_classified.csv
• run_summary.json
• analysis tables and publication figures under full-corpus output folders

Notes

• Some output directories are large and may change frequently between runs.
• Keep run metadata (run_summary.json) for reproducibility and audit trails.
• Script naming reflects research-stage workflows; prefer documenting run parameters in commit messages.

License

Add the preferred license for publication/distribution before sharing publicly.