Tag Me Maybe 1.0

Introduction to TMM

Welcome to Tag Me Maybe - an environment for creating, analysing and presenting linguistic datasets. The application requires no installation and is fully accessible in a web browser.

The application consists of the following modules:

AnnotationCreate and edit annotations together with the relations established between them. ReaderConveniently view a dataset together with its annotations. ExplorerFind the object you are interested in: text, an annotation or a relation. StatisticsAnalyse the dataset using a rich collection of statistics. GraphsVisualise data using graphs. PresentationPresent data in an accessible visual form. Publication versionCheck your dataset and prepare a version for publication.
Annotation · 01

Importing data

You can begin working in TMM with the sample dataset available on the home page or import your own material.

How to start

  1. Download the sample dataset or prepare your own data.
  2. Open TMM and enter the Annotation module.
  3. Use Load and select the appropriate file.
  4. After the data are loaded, you can begin annotation.
Sample data

Start with the example dataset

You can download the sample JSON dataset from the home page and then open it in the Annotation module.

Sample data tile on the TMM home page English TMM home page with the Sample data tile
On the home page, use the Sample data tile and the Download button.
Your own data

Load data

In Annotation, open the Load menu in the bottom toolbar and choose whether you want to import a table or a TMM JSON project.

English Load menu in Annotation

Table import

Use XLSX, CSV or TSV for segment-based source material. A minimal table may contain segment text, with optional IDs and speaker information.

Templates: XLSX · CSV · TSV

JSON project

Use JSON to restore a complete TMM state, including segments, annotations, relations, tag settings and other project data.

💡 Safety: TMM validates imported data before replacing the currently loaded project.

Annotation · 02

Annotation basics

In TMM, the text is divided into segments: basic textual units visible in the left column. On their basis, you can create one or more annotations displayed on the right.

A segment is a fragment of the source text, while an annotation (tags + relations) is information about the text.

To create a tag:

  1. Click a segment in the left column to select it. It will be subtly highlighted.
  2. Choose an existing tag name or type your own.
  3. Add the annotation content.
  4. Confirm by clicking ➕ Add.
Annotation view with highlighted elements needed to create a tag
The screenshot highlights the key elements of the workflow: selecting a segment, choosing a tag, adding annotation content and confirming with Add.
Result of creating an annotation: the tag and its content appear in the right column
After you confirm the operation, the new annotation appears in the right column together with its tag and content.

💡 Tip: Use multiple annotations when one segment performs several functions or contains analytically distinct spans.

Annotation · 03

Annotation hierarchy

In TMM you can define hierarchy between annotations. To do this, create at least two annotations within a single segment, then click the tree icon and set the appropriate dependency.

Annotation view with several annotations in one segment and the tree icon used to edit hierarchy
First create at least two annotations within one segment, then click the tree icon.
Hierarchy editor opened for one segment, before assigning parent-child relations
The hierarchy editor opens and shows all annotations from the selected segment.
Hierarchy editor after assigning a parent-child relation between annotations
Set the parent-child relation and save the changes.
Annotation view showing the saved hierarchy directly in the annotation panel
After saving, the hierarchy becomes visible directly in the annotation panel.
Annotation · 04

Macrostructures

Macrostructures group annotations into larger discourse units. They are useful when the object of analysis extends beyond a single segment. A macrostructure may be discontinuous and may also be nested within another macrostructure.

Macrostructures menu with options for creating a macrostructure and managing macrostructure relations
Open the Macrostructures menu and choose Create macrostructure.
Annotation selection step when creating a new macrostructure
Select the annotations that should belong to the macrostructure, then click Next.
Create macrostructure dialog with type, name, notes and Create button
Enter the macrostructure type and, optionally, a name and analytical notes. Then click Create.
Annotation panel after creating a macrostructure, with macrostructure label, badges and connector line
After creation, the macrostructure becomes visible in the annotation panel and assigned annotations receive its badge.
Annotation · 05

Creating relations

TMM supports two types of relations: (a) between annotations and (b) between macrostructures.

Relations connect two analytical objects: annotations or macrostructures.

To create a relation between annotations:

  1. Right-click the tag pill from which you want the relation to start.
  2. Choose Add relation.
  3. Click the target annotation that should be connected.
  4. Choose a relation from the available list or add your own.
Context menu for an annotation with Add relation highlighted
Right-click the source tag and choose Add relation.
Relation mode with the target annotation indicated
In relation mode, click the target annotation you want to connect.
Relation picker with an existing relation type and an option to add a new one
Choose an existing relation type or add a new one.
Annotation panel showing a created relation between two annotations
After saving, the relation appears in the annotation panel.
Annotation panel showing multiple relations connected to one annotation
One annotation can participate in multiple relations.

To create a relation between macrostructures:

  1. Open the source macrostructure.
  2. Choose Relation.
  3. Select the target macrostructure, complete the form and confirm.
Macrostructure side panel with the Relation button highlighted
Open a macrostructure and click Relation in its side panel.
Macrostructure relation mode with the target macrostructure highlighted and a search field visible
Select the target macrostructure directly or search for it by M# or name.
Create macrostructure relation dialog with fields for relation type, class, directionality and notes
Fill in the relation form and click Save.
Annotation panel showing a created relation between two macrostructures
After saving, the macrostructure relation becomes visible in the annotation panel.
Review and analysis · 01

Reader

The Reader lets you view annotated material in a slightly different way than in the Annotation module. It includes various filters that let you focus on selected objects.

Reader view showing filters, annotations and visible relations in context
The Reader brings together material context, annotation layers, filters and visible relations in one view.
  • Use filters and layer switches to focus on selected relation or macrostructure types.
  • Switch between source-only and bilingual reading when translations are available.
  • Follow the recording to keep the current segment aligned with the media timeline.
Review and analysis · 02

Explorer

Explorer is an advanced search workspace. It lets you precisely search and browse segments, tags and relations.

You can switch between different browsing modes and filter results, for example by tag, speaker, segment, tag content or metadata.

Open Explorer
Explorer view with browsing modes, filters, results and details panel
Explorer combines browsing modes, filters, sortable results and a details panel in one workspace.
Review and analysis · 03

Statistics

The Statistics tab provides a quantitative view of the dataset.

Statistics include both basic information about the dataset and detailed information about segments, tags, relations and speakers.

Open Statistics
Statistics Overview tab showing dataset overview cards, general tags, annotation profile and macrostructure profile
The Overview tab summarises the dataset, annotation profile, general tags and macrostructure indicators.
Relations Structures tab showing relation structure shapes and directional patterns
The Relations → Structures view shows the most common relation shapes and directional patterns in the dataset.
Tag sequences view showing filters and a flow diagram of tag transitions
The Tag sequences analysis helps trace which tags tend to follow one another in the ordered flow of annotations.
Review and analysis · 04

Graphs

The Graphs tab shows annotations and relations in a graph-based form.

It offers a more user-friendly view of relations between tags.

It is especially useful when you want to see the network of connections between tags. Filters are available to configure the relation view as needed. It is worth using when relations between tags are dense.

Open Graphs
Graphs view showing one inference relation between a premise and a conclusion
The Graphs view can display a selected relation as a clear graph connecting its source and target annotations.
Graphs view showing the selected inference relation together with an additional related relation
You can also expand the context to see related annotations and additional connections around the selected relation.
Presentation and data · 01

Presentation

Presentation mode is designed for showing the dataset to an audience. It uses larger typography, controlled layers and layouts intended for a screen or projector.

Open Presentation
Presentation mode showing the debate reading layout with large segment text, a media panel on the left and an analysis panel on the right
Presentation mode combines large reading typography, media context and a structured analysis panel in a projector-friendly layout.
Presentation and data · 02

Export and publication version

Export

Export as a JSON file

An exported JSON file is also a complete save of the project. Create exports as often as possible; they are an insurance policy in case browser storage becomes corrupted.

Export JSON button in the Annotation module
Use the Export JSON button in Annotation to save the complete project as a JSON file.
Prepare dataset

Prepare a publication version

Using this module, you can decide which elements you want to remove from your dataset.

Dataset release tools module
The dataset release module helps you load a file, inspect its structure and prepare a clean publication-ready release.
Open Publication Version tools
Presentation and data · 03

Good practices

  • Save your work regularly.
  • Use one of the available LLMs if you want to prepare your own files quickly; they should handle this task without major difficulty.
  • When importing raw text from .xlsx, make sure the file has the expected structure.
  • Avoid having more than one TMM window open at the same time.
  • It is worth using modules other than Annotation mode, as they can help reveal different kinds of annotation problems.
  • A single segment can contain multiple annotations. This is useful when one utterance performs more than one function.
  • TMM works best on monitors of at least 17 inches.