📊 Adapt visualization scripts for ABM data #16

🐞 Problem Description

The current visualization scripts (e.g., adoption_heatmap_generator) are designed to work with Montecarlo simulation data (data/montecarlo). They are not compatible with ABM simulation data (data/abm), due to differences in CSV format and structure.

Example ABM CSV snippet:

scenario,step,agent_id,type,trust,income,adoption,grid_id
high_trust_policy,1,0,ConsumerAgent,0.9123,,0.0,30
high_trust_policy,1,1,ConsumerAgent,1.0,,0.0,31

⚠️ Issues expected:

• Column names and structure differ from Montecarlo CSVs.
• Scripts may fail when parsing ABM data.
• Aggregation and plotting logic may need adjustments.

🎯 Goal

Make the visualization scripts compatible with ABM data, so that we can generate:

• 🖼️ PNG plots (e.g., adoption heatmaps)
• 📄 LaTeX tables summarizing metrics per scenario/step

🛠️ Tasks

1. 🧩 Implement an ABM CSV adapter to convert ABM CSVs to a format compatible with existing visualization scripts.
2. ⚙️ Update script CLI to accept --data-dir pointing to ABM data directories.
3. 📊 Modify aggregation logic if needed (per scenario/step/agent).
4. 🎨 Ensure output plots and tables are consistent with Montecarlo reports.
5. 🧪 Test scripts on sample ABM runs (data/abm_experiments/run_seed_5678).

✅ Acceptance Criteria

• ✅ Visualization scripts run successfully on ABM data.
• ✅ Generated PNGs and LaTeX tables match the intended metrics.
• ✅ No breaking changes for Montecarlo data processing.