Temporal-network user guide

This guide explains how to build temporal networks, choose measures and interpret outputs with tnet_analysis 1.0.2. Start with small examples before scaling to many nodes, frames or subjects.

The workflow is: signals (optional) → connectivity (optional) → binary network frames → measures. Synthetic networks can enter at the network stage. All examples use generated data and need no external dataset.

Simachew Mengiste (with Demian Battaglia)

Offline reading: Download the PDF user guide.

• Installation and first analysis
  • Install
  • First network
  • Build and browse
• Inputs and time parameters
  • Shapes and conventions
  • Window and lag
  • Hops and latency units
  • Lifetime and search limits
• Signals, FC and null models
  • Methods versus tags
  • Complete signal workflow
  • Thresholding
  • All supported analysis tags
• Choosing and interpreting measures
  • Distance and integration
  • Return latency and retention
  • Dynamism and memory
  • Clustering, partners and communities
• Output layouts and summaries
  • Distance matrices
  • Distance vectors
  • Dynamism order
  • Node-score summaries
  • NaN, infinity and padding
  • Aliases
• Parallel computation and troubleshooting
  • Start small
  • Process-worker example
  • Scratch and checkpointing
  • Common problems
• API reference
  • Connectivity
  • Preparation
  • Surrogates
  • Distance and return
  • Dynamism and memory
  • Structure
  • Modularity
  • Network generators

The toolbox is distributed under the MIT license. Docstrings specify accepted inputs; this guide adds workflow and interpretation advice.