examples

TRACES (Time-series Relationship Analysis with Comprehensive Evaluation Suite)

---

TRACES Example Outputs README

This directory contains example outputs generated using the TRACES analytical framework. All examples are generated using the primary sample dataset TRACES_sample_52x10_dataset_A1.xlsx.

---

TRACES/ ├── data │   └── examples │   ├── TRACES_sample_52x10_dataset_A1.xlsx # Link to reference source data in /data/examples/ │   └── TRACES_sample_52x10_dataset_B1.xlsx # bonus time-series dataset to try, example outputs not provided ├── docs │   ├── ComponentFlexGuide.md │   ├── Formulae.md │   ├── OperationalGuide.md │   └── examples │   ├── README.md # THIS FILE │   ├── outputs │   │   ├── step1_setup.txt # Initial data loading results │   │   ├── step2_correlations.txt # Core correlation analysis │   │   ├── step3_advanced.txt # Advanced correlation methods │   │   ├── step4_framework.txt # Analysis framework results │   │   └── step6_full_analysis.txt # Complete dataset analysis │   └── visualizations │   ├── ccf_analysis.png # CCF Analysis scatter plot │   ├── correlation_comparison.png # Method comparison bar plot │   ├── method_performance.png # Performance by relationship type │   └── relationship_matrix.png # Confidence score matrix

---

TRACES Example Outputs

This directory contains example outputs generated using the TRACES analytical framework. All examples are generated using the primary sample dataset TRACES_sample_52x10_dataset_A1.xlsx.

Source Dataset Reference

Primary dataset: /data/examples/TRACES_sample_52x10_dataset_A1.xlsx

• 52 time points (weekly intervals)
• 11 columns total:
  • Time column
  • 10 time series (Label_1 through Label_10)
• Used to generate all example outputs shown here

Note: An additional sample dataset (TRACES_sample_52x10_dataset_B1.xlsx) is provided in the same directory for users to practice with different time series data.

Directory Contents

Output Files (/outputs/)

Step-by-step analysis results from the TRACES notebook:

1. step1_setup.txt - Initial data loading and validation
2. step2_correlations.txt - Basic correlation analysis
3. step3_advanced.txt - Advanced correlation methods
4. step4_framework.txt - Analysis framework results
5. step6_full_analysis.txt - Complete dataset analysis

Visualizations (/visualizations/)

Four key visualization outputs:

1. ccf_analysis.png - Cross-correlation function analysis scatter plot

   • Shows relationship between maximum correlation and optimal lag
   • Color indicates confidence score

2. correlation_comparison.png - Method comparison bar plot

   • Compares Pearson, Spearman, and Kendall methods
   • Shows relative performance across pairs

3. method_performance.png - Performance by relationship type

   • Distribution of best methods
   • Breakdown by relationship classification

4. relationship_matrix.png - Confidence score matrix

   • Heatmap of relationship types
   • Color intensity shows confidence level

Key Analysis Results

1. Relationship Types Found

   • Complex: 33 pairs
   • Lagged: 7 pairs
   • Linear: 4 pairs
   • Non-linear: 1 pair

2. Correlation Strength Distribution

   • Strong correlations (>0.7): 15 pairs
   • Moderate correlations (0.3-0.7): 26 pairs
   • Weak correlations (<0.3): 4 pairs

3. Method Performance

   • Primary methods: CCF+Spearman (complex relationships), Pearson (linear)
   • Overall mean confidence score: 0.582
   • Highest confidence: 0.999 (Label_3 vs Label_10)

Usage Notes

• All outputs are reproducible using the primary sample dataset
• Examples exhibit the dynamic range of TRACES capabilities
• Text outputs provide detailed analysis at each step
• Visualizations show key relationships and patterns
• Reference the OperationalGuide.md for execution details