Update FRFR tutorial with improved layout and all features

- Reorganize to 2 rows × 3 columns for Figure 1 (PFR, Lag-CRP, SPC)
- Separate Figure 2 for memory fingerprints (early vs late)
- Use crack() to split into early (lists 0-7) and late (lists 8-15)
- Display all 7 features in fingerprint analysis

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Author: jeremymanning

examples/plot_frfr_data.py

@@ -26,9 +26,9 @@
 and lists 9-16 are considered "late" lists.
 
 We'll analyze recall performance using:
-1. Serial Position Curve (SPC) - recall probability by encoding position
-2. Probability of First Recall (PFR) - probability of recalling each position first
-3. Lag-CRP - conditional recall probability by temporal lag
+1. Probability of First Recall (PFR) - probability of recalling each position first
+2. Lag-CRP - conditional recall probability by temporal lag
+3. Serial Position Curve (SPC) - recall probability by encoding position
 4. Memory Fingerprint - clustering by multiple features
 
 Reference:
@@ -73,70 +73,135 @@
 for cond, count in sorted(condition_counts.items()):
     print(f"  {cond}: {count}")
 
-# Build per-subject listgroups: early (lists 0-7) vs late (lists 8-15)
-# Each subject has their own listgroup since we want to compare early vs late
-# within each condition
-listgroup = []
-for subj_idx in range(egg.n_subjects):
-    subj_listgroup = []
-    for list_idx in range(egg.n_lists):
-        if list_idx < 8:
-            subj_listgroup.append('early')
-        else:
-            subj_listgroup.append('late')
-    listgroup.append(subj_listgroup)
-
-# Create a listgroup for averaging all lists together (for fingerprint)
-listgroup_average = ['average'] * egg.n_lists
-
-# Create figure with 2x2 subplots
-fig, axes = plt.subplots(2, 2, figsize=(14, 12))
-
-# 1. Serial Position Curve - by condition, colored by early/late
-print("\nAnalyzing Serial Position Curves...")
-spc = egg.analyze('spc', listgroup=listgroup)
-spc.plot(ax=axes[0, 0], subjgroup=subjgroup, plot_type='subject', legend=True)
-axes[0, 0].set_title('Serial Position Curve by Condition (Early vs Late)')
-axes[0, 0].set_xlabel('Serial Position')
-axes[0, 0].set_ylabel('Recall Probability')
-axes[0, 0].set_ylim([0, 1])
-# Move legend outside plot
-axes[0, 0].legend(loc='upper right', fontsize=7, ncol=2)
-
-# 2. Probability of First Recall - by condition, early/late
-print("Analyzing Probability of First Recall...")
-pfr = egg.analyze('pfr', listgroup=listgroup)
-pfr.plot(ax=axes[0, 1], subjgroup=subjgroup, plot_type='subject', legend=False)
-axes[0, 1].set_title('Probability of First Recall by Condition')
-axes[0, 1].set_xlabel('Serial Position')
-axes[0, 1].set_ylabel('Probability')
-axes[0, 1].set_ylim([0, 0.25])
-
-# 3. Lag-CRP - by condition, early/late
-print("Analyzing Lag-CRP...")
-lagcrp = egg.analyze('lagcrp', listgroup=listgroup)
-lagcrp.plot(ax=axes[1, 0], subjgroup=subjgroup, plot_type='subject', legend=False)
-axes[1, 0].set_title('Lag-CRP by Condition')
-axes[1, 0].set_xlabel('Lag')
-axes[1, 0].set_ylabel('Conditional Recall Probability')
-axes[1, 0].set_xlim([-10, 10])
-axes[1, 0].axvline(x=0, color='gray', linestyle='--', alpha=0.5)
-
-# 4. Memory Fingerprint - by available features
-# Note: color and location are list-type features that require special handling
-print("Analyzing Memory Fingerprints...")
-fingerprint_features = ['category', 'size', 'wordLength', 'firstLetter', 'temporal']
-fingerprint = egg.analyze('fingerprint', features=fingerprint_features,
-                          listgroup=listgroup_average)
-fingerprint.plot(ax=axes[1, 1], subjgroup=subjgroup, plot_type='subject',
-                 title='Memory Fingerprint by Condition', ylim=[0, 1])
-axes[1, 1].set_xlabel('Feature')
-axes[1, 1].set_ylabel('Clustering Score')
-# No legend here since we already have one in SPC plot
+# Split egg into early (lists 0-7) and late (lists 8-15) lists
+print("\nSplitting data into early and late lists...")
+egg_early = egg.crack(lists=list(range(8)))
+egg_late = egg.crack(lists=list(range(8, 16)))
+
+print(f"Early lists egg: {egg_early.n_subjects} subjects, {egg_early.n_lists} lists")
+print(f"Late lists egg: {egg_late.n_subjects} subjects, {egg_late.n_lists} lists")
+
+# Create listgroup for averaging across lists within each split
+listgroup_early = ['average'] * egg_early.n_lists
+listgroup_late = ['average'] * egg_late.n_lists
+
+# ============================================================================
+# Figure 1: PFR, Lag-CRP, SPC for Early (top) and Late (bottom) lists
+# ============================================================================
+print("\n" + "=" * 60)
+print("Creating Figure 1: PFR, Lag-CRP, SPC by condition")
+print("=" * 60)
+
+fig1, axes1 = plt.subplots(2, 3, figsize=(18, 10))
+
+# --- Top row: Early lists ---
+print("\nAnalyzing early lists...")
+
+# PFR - Early
+print("  Computing PFR (early)...")
+pfr_early = egg_early.analyze('pfr', listgroup=listgroup_early)
+pfr_early.plot(ax=axes1[0, 0], subjgroup=subjgroup, plot_type='subject', legend=True)
+axes1[0, 0].set_title('Probability of First Recall (Early Lists)')
+axes1[0, 0].set_xlabel('Serial Position')
+axes1[0, 0].set_ylabel('Probability')
+axes1[0, 0].set_ylim([0, 0.3])
+axes1[0, 0].legend(loc='upper right', fontsize=6, ncol=2)
+
+# Lag-CRP - Early
+print("  Computing Lag-CRP (early)...")
+lagcrp_early = egg_early.analyze('lagcrp', listgroup=listgroup_early)
+lagcrp_early.plot(ax=axes1[0, 1], subjgroup=subjgroup, plot_type='subject', legend=False)
+axes1[0, 1].set_title('Lag-CRP (Early Lists)')
+axes1[0, 1].set_xlabel('Lag')
+axes1[0, 1].set_ylabel('Conditional Recall Probability')
+axes1[0, 1].set_xlim([-10, 10])
+axes1[0, 1].axvline(x=0, color='gray', linestyle='--', alpha=0.5)
+
+# SPC - Early
+print("  Computing SPC (early)...")
+spc_early = egg_early.analyze('spc', listgroup=listgroup_early)
+spc_early.plot(ax=axes1[0, 2], subjgroup=subjgroup, plot_type='subject', legend=False)
+axes1[0, 2].set_title('Serial Position Curve (Early Lists)')
+axes1[0, 2].set_xlabel('Serial Position')
+axes1[0, 2].set_ylabel('Recall Probability')
+axes1[0, 2].set_ylim([0, 1])
+
+# --- Bottom row: Late lists ---
+print("\nAnalyzing late lists...")
+
+# PFR - Late
+print("  Computing PFR (late)...")
+pfr_late = egg_late.analyze('pfr', listgroup=listgroup_late)
+pfr_late.plot(ax=axes1[1, 0], subjgroup=subjgroup, plot_type='subject', legend=False)
+axes1[1, 0].set_title('Probability of First Recall (Late Lists)')
+axes1[1, 0].set_xlabel('Serial Position')
+axes1[1, 0].set_ylabel('Probability')
+axes1[1, 0].set_ylim([0, 0.3])
+
+# Lag-CRP - Late
+print("  Computing Lag-CRP (late)...")
+lagcrp_late = egg_late.analyze('lagcrp', listgroup=listgroup_late)
+lagcrp_late.plot(ax=axes1[1, 1], subjgroup=subjgroup, plot_type='subject', legend=False)
+axes1[1, 1].set_title('Lag-CRP (Late Lists)')
+axes1[1, 1].set_xlabel('Lag')
+axes1[1, 1].set_ylabel('Conditional Recall Probability')
+axes1[1, 1].set_xlim([-10, 10])
+axes1[1, 1].axvline(x=0, color='gray', linestyle='--', alpha=0.5)
+
+# SPC - Late
+print("  Computing SPC (late)...")
+spc_late = egg_late.analyze('spc', listgroup=listgroup_late)
+spc_late.plot(ax=axes1[1, 2], subjgroup=subjgroup, plot_type='subject', legend=False)
+axes1[1, 2].set_title('Serial Position Curve (Late Lists)')
+axes1[1, 2].set_xlabel('Serial Position')
+axes1[1, 2].set_ylabel('Recall Probability')
+axes1[1, 2].set_ylim([0, 1])
 
 plt.tight_layout()
-plt.suptitle('Feature-Rich Free Recall (FRFR) Dataset Analysis', y=1.02, fontsize=14)
-plt.savefig('frfr_analysis.png', dpi=150, bbox_inches='tight')
+fig1.suptitle('FRFR Dataset: Recall Analyses by Condition (Early vs Late Lists)',
+              y=1.02, fontsize=14)
+plt.savefig('frfr_recall_analysis.png', dpi=150, bbox_inches='tight')
+print("\nSaved Figure 1 to frfr_recall_analysis.png")
+
+# ============================================================================
+# Figure 2: Memory Fingerprints for Early (left) and Late (right) lists
+# ============================================================================
+print("\n" + "=" * 60)
+print("Creating Figure 2: Memory Fingerprints by condition")
+print("=" * 60)
+
+fig2, axes2 = plt.subplots(1, 2, figsize=(14, 6))
+
+# Features for fingerprint analysis
+fingerprint_features = ['category', 'size', 'color', 'location',
+                        'wordLength', 'firstLetter', 'temporal']
+
+# Fingerprint - Early lists
+print("\nComputing fingerprints (early lists)...")
+fp_early = egg_early.analyze('fingerprint', features=fingerprint_features,
+                             listgroup=listgroup_early)
+fp_early.plot(ax=axes2[0], subjgroup=subjgroup, plot_type='subject',
+              title='Memory Fingerprint (Early Lists)', ylim=[0, 1])
+axes2[0].set_xlabel('Feature')
+axes2[0].set_ylabel('Clustering Score')
+axes2[0].legend(loc='upper right', fontsize=7, ncol=2)
+
+# Fingerprint - Late lists
+print("Computing fingerprints (late lists)...")
+fp_late = egg_late.analyze('fingerprint', features=fingerprint_features,
+                           listgroup=listgroup_late)
+fp_late.plot(ax=axes2[1], subjgroup=subjgroup, plot_type='subject',
+             title='Memory Fingerprint (Late Lists)', ylim=[0, 1])
+axes2[1].set_xlabel('Feature')
+axes2[1].set_ylabel('Clustering Score')
+# No legend on second plot since first has it
+
+plt.tight_layout()
+fig2.suptitle('FRFR Dataset: Memory Fingerprints by Condition (Early vs Late Lists)',
+              y=1.02, fontsize=14)
+plt.savefig('frfr_fingerprint_analysis.png', dpi=150, bbox_inches='tight')
+print("Saved Figure 2 to frfr_fingerprint_analysis.png")
+
 plt.show()
 
-print("\nAnalysis complete! Saved plot to frfr_analysis.png")
+print("\nAnalysis complete!")