JTNA v1.7.1 — a step-by-step, point-and-click walkthrough of the JTNA jamovi module.

Scope. Every conceptual explanation in this tutorial is drawn from the reference R-package tutorial: Saqr, M., & López-Pernas, S. (2026). An Updated Comprehensive Tutorial on Transition Network Analysis (TNA) — sonsoleslp.github.io/posts/tna-tutorial. JTNA is a jamovi wrapper for the tna R package (Tikka, López-Pernas, & Saqr, 2025), so the behaviour of each analysis option is the behaviour described in that tutorial. Where JTNA exposes features that fall outside that tutorial’s scope (for example Group TNA comparisons or sequence clustering), we describe what each control does at the widget level and point you to the dedicated companion tutorials — tna-group and tna-clustering — rather than attempting to re-explain those methods here.

How this tutorial is laid out. Each top-level section is collapsible — click its heading, or use the Expand / Collapse all pill at the bottom-right, to hide sections you don’t need. The floating table of contents on the left keeps you oriented; clicking a TOC entry scrolls to and opens the target section automatically.

1. What is Transition Network Analysis?

Transition Network Analysis (TNA) is a framework for modelling sequential processes as weighted directed networks. It combines the temporal resolution of stochastic process mining with the structural analytic capacity of graph theory (Saqr, López-Pernas, Törmänen, et al., 2025). The framework supports several model types — first-order Markov models, frequency-based transition models, co-occurrence models, and attention-based models — each suited to different data characteristics and research questions.

What distinguishes TNA from earlier approaches to sequential and process data is that every level of analysis is subject to confirmatory testing:

  • Edge significance is established through bootstrapping.
  • Centrality stability is quantified through case-dropping with correlation stability coefficients.
  • Group differences are evaluated through permutation testing that produces edge-level p-values, effect sizes, and corrections for multiple comparisons.

In the network, nodes are the unique states or actions, arrows are the transitions from one state to another (thicker arrows = higher transition probability), and the coloured rim around each node shows the initial probability of starting in that state.

The framework is implemented in the tna R package (Tikka, López-Pernas, & Saqr, 2025), with companion interfaces in jamovi (this module, JTNA) and a Shiny web application.

2. Installation and Setup

2.1 Install jamovi

Download jamovi from jamovi.org and install it for your operating system.

The jamovi.org download page.
The jamovi.org download page.

2.2 Install JTNA

Option A — jamovi library (recommended):

  1. Open jamovi.
  2. Click the Modules tab (the + icon, top-right).
  3. Click jamovi library.
  4. Search for JTNA and click Install.

Option B — sideload the .jmo file:

  1. Download JTNA_1.7.1.jmo from the JTNA GitHub repository.
  2. In jamovi, click Modules → Sideload.
  3. Open the downloaded .jmo file.
The Modules panel after JTNA has been installed.
The Modules panel after JTNA has been installed.

2.3 Load the example dataset

  1. Click the hamburger menu (top-left).
  2. Open → Data Library.
  3. Under the JTNA section, pick Regulation_long.
jamovi’s Data Library with the Regulation_long entry selected.
jamovi’s Data Library with the Regulation_long entry selected.

3. The example dataset

Throughout this tutorial we use the bundled Regulation_long dataset, a long-format event log of coded collaborative-regulation behaviours.

Property Value
Rows 13,767 events
Actors 2,000 individuals
Actions (9 states) adapt, cohesion, consensus, coregulate, discuss, emotion, monitor, plan, synthesis
Groups (3) A, B, C
Columns Actor, Achievers (High/Low), Group (A/B/C), Time_Posix, Action, Time_Unix

Each row is one event: one actor performed one action at one time. The data is already in long format — exactly the shape JTNA (and the underlying prepare_data() function) expects.

jamovi’s spreadsheet view showing the six columns of Regulation_long.
jamovi’s spreadsheet view showing the six columns of Regulation_long.

4. Part 1 — TNA module

Open Analyses → JTNA → TNA.

The left-hand options panel is organised top-to-bottom into the following sections, which this tutorial walks through in order:

  1. Variables (Action / Actor / Time / Order)
  2. Basic Options (Type, Matrix, Scaling, Threshold)
  3. ━━ Network Analysis ━━ — Visualization Settings, Centrality Analysis, Edge Betweenness, Community Detection, Clique Analysis
  4. ━━ Confirmatory Testing ━━ — Bootstrap Analysis, Centrality Stability
  5. ━━ Sequence Analysis ━━ — Sequence Analysis, Pattern Discovery, Sequence Indices

4.1 Assigning variables

The variable supplier at the top has four target slots.

Slot Column to drag in Notes
Action Action Required. Each unique value becomes a node in the network.
Actor Actor Optional but strongly recommended — it tells JTNA to build one sequence per person. Without it the function treats the whole data frame as one long sequence, chaining actor A’s last event directly into actor B’s first event, which makes no substantive sense.
Time Time_Unix Optional. Within each actor, events are sorted by timestamp; also, two consecutive events from the same actor more than 15 minutes apart (the default) are treated as belonging to different sequences.
Order (leave empty) Optional. Use this instead of Time when you only have a step counter, not a timestamp. If both are supplied, Time sorts first and Order breaks ties.

Any column in the data that is not assigned to one of these slots (for example Achievers, Group) is preserved as metadata and is available later as a grouping variable in Group TNA.

Drag Action → Action, Actor → Actor, Time_Unix → Time.

TNA options panel with Action/Actor/Time_Unix assigned.
TNA options panel with Action/Actor/Time_Unix assigned.

As soon as Action is assigned, JTNA builds the model and starts producing output on the right.

4.2 Basic Options

Type (model type)

Option Description
Relative (default) First-order Markov transition-probability model (default). Rows of the transition matrix sum to 1; each cell [i, j] is the probability of transitioning from state i to state j.
Frequency Frequency-based transition model (raw counts)
Attention Attention-based model. The decay rate is controlled by the Lambda field (only enabled when Attention is selected).

Matrix (checkbox)

When ticked, JTNA shows the full transition matrix. Each cell [i, j] is the probability of transitioning from state i to state j, and rows sum to 1.

The 9×9 transition matrix produced for Regulation_long.
The 9×9 transition matrix produced for Regulation_long.

Scaling

Four options: No Scaling (default), MinMax, Max, Rank. We recommend leaving this at No Scaling. See the tna R-package documentation if you need to apply a non-default scaling.

Parameters → Threshold (default: 900)

This is the session-splitting threshold in seconds — the time_threshold argument of prepare_data() in the reference tutorial. Its default (900 = 15 minutes) means that if two consecutive events from the same actor are more than 15 minutes apart, JTNA treats them as belonging to different sequences. A transition from the last event before a break to the first event after a break is not a real transition — there is a gap in between where the process stopped. Adjust this to match your data: a chat corpus might use 2 minutes (120), a weekly longitudinal study might use one day (86400).

Basic Options section showing Type, Matrix, Scaling, and Threshold.
Basic Options section showing Type, Matrix, Scaling, and Threshold.

4.3 Visualization Settings

Display Options

Checkbox What it shows Reference-tutorial equivalent
TNA plot (on by default) The main transition-network plot. plot(model, minimum = ..., cut = ...)
Histogram Distribution of all transition probabilities. Useful for choosing a pruning threshold. hist(model)
Frequencies plot Bar chart of how often each state appears in the data (raw frequency, not transition probability). plot_frequencies(model)
Mosaic plot Mosaic-plot view of transition frequencies. Not covered in the reference tutorial — use only if you specifically want a mosaic visualisation.
Transition network plot for Regulation_long.
Transition network plot for Regulation_long.
Histogram of transition probabilities for Regulation_long.
Histogram of transition probabilities for Regulation_long.
Frequencies plot for Regulation_long.
Frequencies plot for Regulation_long.

Plot Parameters

The two parameters the reference tutorial explains explicitly are Cut value and Minimum value. They map onto plot(model, minimum = , cut = ) as follows:

Parameter Default Meaning (from reference tutorial)
Minimum value 0.05 Edges with weight below this value are hidden entirely but are retained in all analysis.
Cut value 0.1 Edges with weight below this value appear faded/thinner.
Edge label size 1 Size of the numeric label on each arrow (cosmetic).
Node size 8 Size of each node (cosmetic).
Node label size 1 Size of the state-name labels (cosmetic).
Layout Circle Node-placement algorithm. 14 options are available; the reference tutorial uses the default.
Digits 1 Decimal places for mosaic-plot labels.

Important: lowering the Minimum value does not change any analysis — it only affects what appears on the plot. All edges remain in the model and are used by every downstream statistic.

4.4 Centrality Analysis

Centrality measures reduce each state to a single number capturing its structural importance in the network. Different measures answer different questions: which states are the most common destinations (InStrength), the most common origins (OutStrength), the key bridges between other states (Betweenness), or the most influential in spreading activation (Diffusion).

Display Options

  • Table — one row per state, one column per selected measure.
  • Plot — one panel per selected measure.

Measures

Measure Interpretation
OutStrength Total outgoing weight. Frequent origin state.
InStrength Total incoming weight. “Attractor” in the process.
Closeness Overall connectedness.
ClosenessIn / ClosenessOut Reachability from/to other states.
Betweenness Bridge/bottleneck between other states.
BetweennessRSP Robust betweenness via randomized shortest paths.
Clustering Whether neighbours are also connected to each other.
Diffusion Spread of activation through the network.

Parameters

  • Include loops — include self-loops when computing centrality (reference tutorial: loops = TRUE).
  • Normalize values — rescale measures (reference tutorial: normalize = TRUE).
Centrality table and multi-panel centrality plot.
Centrality table and multi-panel centrality plot.

4.5 Edge Betweenness

While node centrality summarises the importance of individual states, edge betweenness identifies the most critical transitions — the connections that serve as bridges between different parts of the process. Transitions with high edge betweenness lie on many shortest paths and therefore play a key structural role.

  • Table — edge betweenness value for every transition.
  • Plot — a network where edge thickness reflects betweenness.

This corresponds to the reference tutorial’s betweenness_network(model) → plot(..., cut = 0.1).

The Plot Parameters (Cut, Minimum value, Edge label size, Node size, Node label size, Layout) behave exactly as in Section 4.3.

Edge-betweenness table and network plot.
Edge-betweenness table and network plot.

4.6 Community Detection

Communities are groups of states that are more densely connected to each other than to the rest of the network. Unlike cliques (which require every pair to be mutually connected), community-detection algorithms use modularity-based criteria to reveal the modular structure and functional subsystems within the transition network.

In the reference tutorial, communities(model) runs seven algorithms and prints how many communities each one detected, together with the assignment of each state to a community under each algorithm.

Methods

Method Logic
leading_eigen Leading eigenvector of the modularity matrix.
walktrap Random walks; nodes visited together belong together.
fast_greedy Greedy modularity optimisation.
spinglass Spin-glass model for fine-grained structure.
infomap Description-length minimisation for directed flow.
label_prop, edge_betweenness Additional algorithms supported by the underlying tna package.

JTNA default: Spinglass. The reference R-tutorial uses leading_eigen by default. Either is fine — compare the assignments across several methods rather than relying on a single algorithm.

Parameters

  • Gamma — resolution parameter used by Spinglass (higher → more, smaller communities).
Community assignment table and coloured community plot.
Community assignment table and coloured community plot.

4.7 Clique Analysis

A clique is a fully connected subnetwork where every state has strong mutual transitions to every other state in the group. Cliques reveal self-reinforcing behavioural loops — for example, a ‘plan → monitor → adapt → plan’ cycle where each state frequently transitions to each of the others. These are substantively interesting because they represent behavioural patterns that sustain themselves: once a learner enters a clique, the transition probabilities keep them cycling within it.

Parameters

Parameter Meaning
Size Number of states in each clique (2 = dyads, 3 = triads, 4 = quads, …).
Threshold Minimum transition probability that every edge in the clique must meet.

Practical guidance from the reference tutorial: lower the threshold as clique size increases, because fully mutual connections become increasingly unlikely. The reference tutorial uses threshold = 0.1 for dyads, 0.05 for triads, 0.03 for quads.

Display

  • Plot — a panel of subnetwork plots, one per detected clique.
Dyad clique plots.
Dyad clique plots.

Triad clique plots.
Triad clique plots.

4.8 Bootstrap Analysis (Confirmatory Testing)

The transition probabilities estimated by tna() are descriptive estimates derived from a particular sample of sequences. Without further validation, there is no basis for determining which of these estimates reflect stable properties of the underlying process and which are artifacts of sampling variability. Bootstrapping addresses this directly by resampling sequences with replacement and reconstructing the transition matrix across a large number of iterations … Edges that do not consistently exceed a defined threshold or that deviate beyond an acceptable consistency range are identified as non-significant. The bootstrap is therefore not an optional diagnostic — it is a confirmatory step that determines whether the network structure warrants substantive interpretation.”

Parameters

JTNA field Default Reference-tutorial argument
Iteration 1000 iter = 1000
Level 0.05 level = 0.05
Method → Stability (default) — method = "stability"
Method → Threshold — method = "threshold"
Lower / Upper (under Method Parameters, Stability) 0.75 / 1.25 consistency_range = c(0.75, 1.25)
Threshold (under Method Parameters, Threshold method) 0.1 edge-weight threshold

Display

  • Plot — network plot showing only edges that survived the bootstrap.
  • Table — bootstrap summary, with one row per edge and the columns below.

Bootstrap-summary columns (reference tutorial)

Column Meaning
from, to Source and target state.
weight Original observed transition probability.
p_value Proportion of inconsistent resamples.
sig Whether the edge is significant at the chosen level.
cr_lower, cr_upper Consistency-range bounds.
ci_lower, ci_upper Bootstrap confidence-interval bounds.

Interpreting the results

  • Significant edges (sig = TRUE) — appeared consistently across bootstrap resamples. These transitions are stable properties of the data and warrant substantive interpretation.
  • Non-significant edges (sig = FALSE) — unstable under resampling. Their presence in the original model may reflect sampling noise rather than a genuine transition pathway.
  • Narrow confidence intervals mean the edge weight is precisely estimated; wide confidence intervals mean substantial uncertainty remains even if the edge reached significance.
  • Practical implication: report and interpret only the bootstrap-validated network. Conclusions drawn from non-validated edges risk being non-replicable.

Table Options

  • Max rows (default 20) / Show all rows — paging controls.
  • Significant only — filter the table to sig = TRUE edges.
Bootstrap summary table (first 20 rows) and the bootstrap-validated network plot.
Bootstrap summary table (first 20 rows) and the bootstrap-validated network plot.

4.9 Centrality Stability (Confirmatory Testing)

Centrality measures are among the most frequently interpreted yet least frequently validated quantities in network-based analyses. The case-dropping bootstrap systematically removes increasing proportions of sequences, recomputes centrality at each step, and quantifies how much the original ranking is preserved — via the Correlation Stability (CS) coefficient (Epskamp, Borsboom, & Fried, 2018). The CS coefficient is the maximum proportion of cases that can be dropped while maintaining a correlation of at least 0.7 with the original centrality ordering.

CS coefficient Interpretation
≥ 0.7 Excellent — centrality rankings are robust and can be interpreted with confidence.
0.5 – 0.7 Acceptable — rankings are reasonably stable but should be interpreted with some caution.
< 0.5 Poor — rankings are unreliable and should not serve as the basis for substantive conclusions.

“A CS coefficient below 0.5 indicates that the centrality ordering is not sufficiently stable … the appropriate response is to collect additional data or to refrain from interpreting centrality rankings altogether.”

Display

  • Stability Table — CS-coefficient for each selected measure.
  • Stability Plot — the average correlation between full-sample and reduced-sample centrality rankings as the drop proportion increases.

Measures

InStrength, OutStrength, Betweenness (the three on by default).

Parameters

  • Iterations (default 100).
  • Threshold (default 0.7) — the correlation target used to compute the CS coefficient.
  • Certainty (default 0.95) — confidence level for the CS estimate.
Centrality-stability table and stability plot.
Centrality-stability table and stability plot.

4.10 Sequence Analysis

The sections above all work with the transition network, where each edge represents how often one state follows another across all sequences. Sequence analysis works with the raw sequences directly, before they get collapsed into a transition matrix. It provides a complementary perspective: the full temporal unfolding of individual sequences and the overall distribution of states across sequence positions.

This corresponds to the reference tutorial’s plot_sequences(prepared_data).

Plot Type

  • Index (default) — each row is one individual sequence, colours represent states at each position. Reveals patterns in sequence length, composition, and ordering that the transition matrix alone cannot capture.
  • Distribution — aggregates across all sequences, showing at each position the proportion (or count) of sequences in each state.

Scale

  • Proportion (default) / Count — y-axis metric for the distribution plot.

Geometry

  • Bar (default) / Area — stacked bars or stacked area chart.

Options

  • Include missing values (default off in TNA, on in Group TNA / Cluster TNA).
  • Tick Interval (default 5) — spacing of x-axis tick marks.
Distribution sequence plot.
Distribution sequence plot.

4.11 Pattern Discovery

This section lets JTNA mine your sequences for recurring sub-sequences (n-grams, gapped patterns, repeated patterns) using the codyna package.

Control Effect
Show patterns table Toggles the table on.
Pattern Type N-gram (consecutive), Gapped (with gaps), Repeated (repeated within an individual).
Min Length / Max Length Length range for N-gram and Repeated patterns.
Min Gap / Max Gap Gap range, only used when Pattern Type = Gapped.
Min Support / Min Count Minimum proportion of actors, and minimum absolute count, required for a pattern to be reported.
Starts with / Ends with / Contains String filters applied to the pattern.
Max rows / Show all rows Table paging.

Refer to the codyna package documentation for the substantive interpretation of these measures.

Pattern Discovery controls.
Pattern Discovery controls.
Example patterns table.
Example patterns table.

4.12 Sequence Indices

Computes per-actor summary statistics about each individual’s sequence (for example entropy and diversity measures).

Control Effect
Show indices table Toggles the table on.
Favorable State Pick one action from the dropdown to compute indices related to time spent in that state.
Omega (default 1) Weighting parameter for the favorable-state indices.
Max rows / Show all rows Table paging.

Refer to the package documentation for formulas and interpretation.

Sequence Indices table.
Sequence Indices table.

5. Part 2 — Group TNA

Open Analyses → JTNA → Group TNA.

Group TNA builds a separate transition network for every level of a grouping variable and adds comparison and permutation-testing tools. In the reference R-tutorial, group comparisons are covered in a dedicated companion tutorial (tna-group). This section walks through the JTNA UI and grounds each shared feature in the reference tutorial; for the group-comparison-specific tools it describes the widgets only and points you to the companion tutorial for the conceptual treatment.

5.1 Assigning variables

Group TNA adds one new slot on top of the four TNA slots:

Slot Column
Action Action
Actor Actor
Time Time_Unix
Order (leave empty)
Group Group

Drag Group into the Group slot. JTNA will detect levels A, B, C and build one network per level.

Group TNA options panel with Group assigned.
Group TNA options panel with Group assigned.

5.2 Per-group outputs

The sections that follow are identical in meaning to the TNA module — they simply produce one result per group.

  • Visualization Settings — Section 4.3 applies, per group.
  • Centrality Analysis — Section 4.4 applies, per group.
  • Community Detection — Section 4.6 applies, per group.
  • Clique Analysis — Section 4.7 applies, per group.
  • Bootstrap Analysis — Section 4.8 applies, per group. Each group’s network is validated independently.
  • Sequence Analysis / Sequence Indices — Sections 4.10 and 4.12 apply, per group.

Absent from Group TNA. Edge Betweenness, Centrality Stability and Pattern Discovery are exposed only in the TNA module — they do not appear in Group TNA’s panel.

Three per-group TNA plots, one each for Group A, B and C.
Three per-group TNA plots, one each for Group A, B and C.

5.3 Permutation Analysis (group-comparison feature)

Permutation testing assesses whether observed differences in transition probabilities between two groups exceed what would be expected under a random reassignment of group labels. This feature is described in detail in the companion tutorial (tna-group) — we describe only the widget here.

JTNA field Default Notes
Iter 1000 Number of random permutations.
Level 0.05 Significance threshold.
Paired off Tick when groups are naturally paired (e.g. pre/post on the same actors).
Text / Plot off Output toggles.
Max rows / Show all rows 20 / off Table paging.
Permutation-test table with edge-level p-values.
Permutation-test table with edge-level p-values.

5.4 Network Difference Plot (group-comparison feature)

Produces a single network plot colouring each edge by which group has the stronger transition. See the companion tutorial for interpretation. The Plot Parameters (Cut, Min value, Edge label size, Node size, Node label size, Layout) behave as described in Section 4.3.

Network difference plot comparing two groups.
Network difference plot comparing two groups.

5.5 Compare Network Properties (group-comparison feature)

Widget-level description only — consult the companion tutorial for methodology.

  • First Network / Second Network — pick two group levels from the dropdowns.
  • Summary Metrics Table — side-by-side global metrics for the two networks.
  • Network Properties Table — edge-level comparison (weights, difference, correlation, distance).
  • Comparison Plot + Plot Type — Heatmap (default), Scatterplot, Weight Density.
  • Scaling Method — None (default), MinMax, Max, Rank, Z-Score, Robust, Log, Log1p, Softmax, Quantile. Controls how edge weights are scaled before the two networks are compared.
Summary metrics table and network-properties table.
Summary metrics table and network-properties table.
Heatmap comparison plot.
Heatmap comparison plot.

5.6 Compare Sequences (group-comparison feature)

Tests whether specific short action sub-sequences differ in frequency between groups. Widget-level description only.

Control Default Notes
Sub Min / Sub Max 2 / 4 Range of sub-sequence lengths to test.
Minimum Frequency 20 Sub-sequences below this total count are ignored.
Correction Method Bonferroni Also available: Holm, Hochberg, BH (FDR), BY, None.
Max rows / Show all rows 20 / off Table paging.
Compare-Sequences table with adjusted p-values.
Compare-Sequences table with adjusted p-values.

6. Part 3 — Cluster TNA

Open Analyses → JTNA → Cluster TNA.

Cluster TNA is for situations where you do not have a pre-defined grouping variable. It clusters actors based on the similarity of their sequences, then builds a separate transition network for each discovered cluster. The underlying clustering methodology is covered in a separate companion tutorial (tna-clustering) — as with Group TNA, this section walks through the JTNA UI and grounds shared features in the reference tutorial, while describing clustering-specific controls only at the widget level.

6.1 Assigning variables

The four standard slots (Action, Actor, Time, Order) — no Group slot, because clusters are discovered from the data.

Cluster TNA options panel.
Cluster TNA options panel.

6.2 Clustering Parameters

The first section in the Cluster TNA panel is Clustering Parameters (expanded by default). This is where Cluster TNA differs from the other two modules.

Control Default Notes
Distance Measure Hamming Also: Optimal String Alignment (OSA), Levenshtein, Longest Common Substring (LCS), Jaccard, Jaro-Winkler.
Clustering Method PAM Also: Ward (D2), Complete, Average, Single.
Number of Clusters (k) 2 Integer, 2–20.
Run Clustering off Must be ticked for clustering to run.

Clustering can be computationally expensive on large datasets, which is why the Run Clustering checkbox has to be ticked explicitly. See the clustering companion tutorial for guidance on picking a distance measure and k.

Clustering Parameters section with Run Clustering ticked.
Clustering Parameters section with Run Clustering ticked.

6.3 Per-cluster outputs

Once clustering completes, everything below is produced per-cluster. These sections behave exactly as in the TNA module:

  • Visualization Settings — Section 4.3.
  • Centrality Analysis — Section 4.4.
  • Community Detection — Section 4.6.
  • Clique Analysis — Section 4.7.
  • Bootstrap Analysis — Section 4.8.
  • Sequence Analysis / Sequence Indices — Sections 4.10 and 4.12.
Two per-cluster network plots for k = 2.
Two per-cluster network plots for k = 2.

6.4 Cluster comparisons

The comparison tools are the same as in Group TNA, with two UI differences:

  • First Network / Second Network are integer inputs (cluster 1, 2, …), not dropdowns.
  • Scaling Method offers a reduced list: None, MinMax, Max, Rank, Z-Score, Robust.

Everything else — Permutation Analysis, Network Difference Plot, Compare Network Properties, Compare Sequences — mirrors Sections 5.3–5.6. See the clustering companion tutorial for interpretation.

Cluster comparison: permutation table.
Cluster comparison: permutation table.
Cluster comparison: network difference plot.
Cluster comparison: network difference plot.
Cluster comparison: heatmap.
Cluster comparison: heatmap.

7. Complete workflow at a glance

JTNA exposes the same end-to-end analysis pipeline that the reference R-tutorial describes:

Step Reference-tutorial R call JTNA equivalent
Prepare data prepare_data(df, action=, actor=, time=, time_threshold=900) Variable slots (Action / Actor / Time / Order) + Parameters → Threshold (900).
Build model tna(prepared) Happens automatically once Action is assigned. Pick a Type.
Visualise plot(model, minimum=0.05, cut=0.1) Visualization Settings → TNA plot (Minimum value = 0.05, Cut value = 0.1).
Inspect distributions hist(model); plot_frequencies(model) Visualization Settings → Histogram; Frequencies plot.
Cliques cliques(model, size=k, threshold=t) Clique Analysis (Size, Threshold).
Node centrality centralities(model) Centrality Analysis.
Edge centrality betweenness_network(model) Edge Betweenness.
Communities communities(model, methods=…) Community Detection (Methods).
Bootstrap bootstrap(model, iter=1000, level=0.05) Bootstrap Analysis (Iteration, Level, Method, Range).
Centrality stability estimate_centrality_stability(model) Centrality Stability.
Raw sequences plot_sequences(prepared) Sequence Analysis.

For group-level and cluster-level comparisons (Permutation Analysis, Compare Network Properties, Compare Sequences, Network Difference Plot), consult the companion tutorials tna-group and tna-clustering — they ground those features in the same way this tutorial grounds the TNA module.

8. Troubleshooting and tips

General

  • No output appears. Ensure the Action variable is assigned and that the data has at least two unique action values.
  • Greyed-out options. Many controls are only enabled when a parent toggle is on (e.g. Plot Parameters require the relevant Plot / Table checkbox). Lambda is only enabled when Type = Attention. Gap fields are only enabled when Pattern Type = Gapped. Mosaic plot requires Type = Frequency.
  • Bootstrap is slow. Reduce Iteration from 1000 to 100 while exploring, then raise it again for the analysis you report.
  • Pick a sensible Minimum value. Tick Histogram first to see the distribution of edge weights, then pick a minimum that removes visual noise without discarding edges you actually care about. Remember: Minimum value only affects the plot — every analytical result uses the full set of edges.
  • Clustering is slow. Large datasets with many actors take longer. Try reducing k or picking a faster distance measure (Hamming or Jaccard).

9. References and citation

To cite JTNA

López-Pernas, S., Girault, D., Tikka, S., & Saqr, M. (2026). JTNA: A Desktop Software for Transition Network Analysis. Proceedings of ICSLE 2025. The Changing Landscape of Smart Learning Environments: Pedagogy, Collaborative Intelligence, and Ethics. LNET.

To cite the TNA methodology and R package

  • Saqr, M., López-Pernas, S., Törmänen, T., Kaliisa, R., Misiejuk, K., & Tikka, S. (2025). Transition Network Analysis: A Novel Framework for Modeling, Visualizing, and Identifying the Temporal Patterns of Learners and Learning Processes. In Proceedings of the 15th International Learning Analytics and Knowledge Conference (LAK ’25) (pp. 351–361). ACM. https://doi.org/10.1145/3706468.3706513
  • Tikka, S., López-Pernas, S., & Saqr, M. (2025). tna: An R Package for Transition Network Analysis. Applied Psychological Measurement. https://doi.org/10.1177/01466216251348840
  • Saqr, M., López-Pernas, S., & Tikka, S. (2025). Mapping Relational Dynamics with Transition Network Analysis: A Primer and Tutorial. In Saqr & López-Pernas (Eds.), Advanced Learning Analytics Methods. Springer. https://lamethods.org/book2/chapters/ch15-tna/ch15-tna.html
  • Saqr, M., López-Pernas, S., & Tikka, S. (2025). Capturing The Breadth and Dynamics of the Temporal Processes with Frequency Transition Network Analysis: A Primer and Tutorial. In Saqr & López-Pernas (Eds.), Advanced Learning Analytics Methods. Springer. https://lamethods.org/book2/chapters/ch16-ftna/ch16-ftna.html
  • López-Pernas, S., Tikka, S., & Saqr, M. (2025). Mining Patterns and Clusters with Transition Network Analysis: A Heterogeneity Approach. In Saqr & López-Pernas (Eds.), Advanced Learning Analytics Methods. Springer. https://lamethods.org/book2/chapters/ch17-tna-clusters/ch17-tna-clusters.html

Companion tutorials (used by this document)

Tutorial written for JTNA v1.7.1. If you are using a different version some options or defaults may differ.