While the standard Obsidian graph view may be interesting for generating pretty visuals, it lacks most of the metrics that make network graphs truly useful. That's why we created the InfraNodus Obsidian graph view plugin. Not only does it provide the most advanced graph insights and analytics, but it also has a beautiful 3D interface, AI capabilities (no OpenAI account needed), and can identify gaps in your knowledge to help you generate new ideas.
This plugin's main advantage is that it can represent not only the connections between your [[wiki links]], but also the concepts contained inside your texts. Therefore, even if you don't use the Obsidian's [[wiki-links]] feature to link concepts together, InfraNodus will do that for you! Of course, you can switch this option off, but we keep it on by default as it's a very powerful feature that sets this plugin apart from any other graph view plugin for Obsidian.
You can use advanced InfraNodus network insights for
- extracting the main topical clusters for your
- individual pages
- selected pages
- folders
- search results
- bookmark groups
- identifying the most influential pages and concepts
- revealing the structural gaps in your ideas, which is useful for
- generating insights
- coming up with new connections
- getting information about trends and emerging topics
This plugin works both on desktop and mobile. It requires an InfraNodus account.
Video Tutorials
Part 1: Introduction to InfraNodus Obsidian graph view
Part 2: Using advanced network science to analyze your Obsidian graphs
InfraNodus Obsidian Graph View Features
Here are some of the main features of the InfraNodus Obsidian graph view plugin:
1. Visualize the main ideas on any page:
- take both mentions [[wiki-links]] and concepts into account (can be adjusted)
- use advanced clustering algorithms and a beautiful 3D visualization module
- show connections between ideas inside your text
2. Identify the main topics and most important ideas
Use advanced network science algorithms to:
- detect clusters of concepts that appear together
- highlight the most relevant concepts (shown bigger on the graph)
- locate original content that belongs to the topics identified
- slice off the top layer ("Reveal Underlying") to show hidden ideas
3. Identify gaps and generate new ideas using the built-in AI
To help you generate new ideas, you can run the built-in AI on:
- the whole graph structure
- selected topical clusters
- selected concepts
- selected statements
- gaps
- trends
- using the built-in AI chat
- InfraNodus Obsidian graph view plugin AI generation ideas
4. Include the backlinks and unlinked mentions into analysis
To help you find hidden connections between your ideas, you can:
- include the snippets from backlink mentions into analysis (adjustable)
- include unmentioned snippets into analysis
- InfraNodus Obsidian backlinks analysis
5. Visualize from any content view: search results, bookmark groups, selected files
No other graph view plugin offers so many visualization options. You can extract the content to visualize from:
- a single file
- selected files
- vault folders
- search result snippets
- pages found in the search results
- bookmark folders
- InfraNodus Obsidian search results visualization
6. Obtain advanced network science metrics on your graph
This plugin has the most comprehensive graph analytics of any Obsidian plugin, including:
- betweenness centrality (bc)
- relative influence
- graph's modularity
- concept's relevance (bc + degree)
- main relations (bigrams)
- graphDot structure for further analysis (e.g. using our Custom GPT)
- InfraNodus graph analytics
7. Navigate through your vault using the graph
You can navigate between pages using the arrow button
User Manual
The best use case for the plugin is to get an overview of the connections between your ideas and to find the gaps between different topics. Here's how you can do that:
- Open your Obsidian vault on any page.
- Click the InfraNodus Graph button (available at the top menu, left menu, and sidebar menu). You can also activate the plugin for any folder: just right-click on it and choose "InfraNodus graph".
- InfraNodus will open a graph visualization of the page's content. The words and the [[wikilinks]] that you use are the nodes and their co-occurrences are the connections between them. Based on network science metrics, the nodes will be ranged by their importance (betweenness centrality) and aligned into groups (topical clusters), which will have the same color. Based on this representation, you can see what are the most important ideas and what topical clusters exist in your document.
- To learn more about the science behind the tool, please, check the [InfraNodus - How it Works](https://infranodus.com/about/how-it-works) page as well as the the peer-reviewed [InfraNodus Whitepaper](https://dl.acm.org/doi/10.1145/3308558.3314123).
- Use the graph visualization to get an overview of what the text is about and see if anything is missing.
- Click on the "Topics" to see the main topics present in the document. You can use the built-in GPT-4 AI to generate a summary of the document in "Topics" > "Summary".
- Click on the terms that seem relevant to you and then click on the "Concepts" > "Context" button to see the context where these terms appear in the document (the actual statement). You can also navigate directly to that statement in Obsidian and visualize a new graph of the document where the statement is contained.
- The most interesting feature: click on the "Gaps" and see the blind holes identified in your content. These are the clusters of ideas that could be better connected. Use the built-in AI to generate interesting research questions that will help you develop the ideas in your notes further.
- Remove the top nodes from the graph (select > hide or the "Reveal Underlying" button on the "Concepts" layer) to reveal underlying ideas and latent topics that are not visible on the surface.
- Reiterate.
- If you find a concept or topic you like, you can also use InfraNodus to navigate to that page. Just click the node in the graph and then click the arrow above to proceed to that page.
- Feel free to export the most interesting excerpts and AI-generated ideas to InfraNodus for later reference and analysis.
Main Interface Features
The plugin is created to promote ecological thinking following the cognitive variability framework. In short, its objective is to help you explore your ideas and generate insights through alternation of scale and intention. The graph interface takes you through the successive stages of zooming in (focusing on specific ideas) to zooming out (getting an overview), focusing (following the existing threads) and exploring (creating your own).
Below you will find information about the main interface features that will help you in this process.
Plugin Views
The plugin has 5 main views:
- general "graph" view (shows the graph + the insights about the structure)
- "topics" view (shows the topics identified)
- the extended view shows additional stats
- "concepts" view (shows the top concepts and their relations)
- the extended view shows additional stats, including
- bigrams
- dotGraph structure (can be interpreted using other tools, e.g. our Custom GPT)
- the extended view shows additional stats, including
- "gaps" view (shows the gaps — topics that could be better connected)
- "trends" view (shows how the topics evolve chronologically + trending topics)
Layer 1: Graph View and Network Structure Insights | Overview
By default, the graph view shows the connections between the [[wiki links]] and concepts retrieved from the files or search results you are analyzing. You can change this behavior in the plugin's settings. The nodes are the [[wiki links]] and the concepts, the edges are the connections.
The bigger are the nodes, the more influential they are (based on the betweenness centrality measure used in network science).
The nodes ([[wiki links]] and concepts) that tend to appear often in the same context will be shown closer to each other on the graph and have the same color (based on Force Atlas layout and modularity algorithms).
The statistics below shows insights about the graph's structure: whether it's biased / focused / diversified / dispersed and a recommendation for optimization (e.g. if it's too biased or focused on one idea the recommendation is to disperse, if it's too dispersed, the recommendation is to connect the clusters together).
Clicking on the recommendation will take you to the layer where you can perform the recommended action.
Layer 2: Topics and Topical Clusters | Zooming In / Exploring
The second layer "Topics," shows the main topical clusters identified. This is useful for getting a visual overview of the content and identifying the main clusters of ideas.
They are based on the [[wiki links]] and concepts that most frequently co-occur in the same paragraph together (this is modifiable in the plugin's settings). You can click those clusters to highlight them in the graph.
When you are in this mode, clicking on the graph will highlight the topic to which the nearest nodes belong.
Once you're in the "topics" layer, you can use the built-in AI features and
- click on the "Context" button to see the statements from the file that belong to all / selected topics
- click on the "Question" button to generate a research question that relates to all / selected topics
- click on the "Idea" button to generate an idea based on all / selected topics
- click on the "Summary" button to summarize all the statements in all / selected topics
If you select a topic, you can click "Locate" button to see where it appears.
Layer 3: Concepts and Relations | Zooming In / Focusing
The third layer shows the main concepts and their relations. This is useful for zooming in and getting into the nuance of text.
You can discover:
- The top concepts (most influential ideas, based on betweenness centrality)
- The top relations (bigrams)
- click on the expand button to see that
- The Graph Dot representation (compressed information about the graph's connectivity to be interpreted by external tools or AI)
- Underlying ideas (click "Reveal Underlying" to hide the top nodes to see what's hiding underneath)
- Related ideas (once you select a few nodes and click the "Link" button, you can see them underneath)
Layer 4: Structural Gaps | Zooming Out / Exploring
The "gaps" layer shows the topical clusters that could be better connected but are not. Trying to link them together will lead you to generate new interesting links in your content and come up with new ideas.
You can load more gaps by clicking "reload".
You can see more information about each gaps (the concepts contained inside) by expanding the view.
Layer 5: Trends and Narrative Timeline Evolution | Zooming Out / Exploring
The "trends" panel shows how the topics evolved over time. It also identifies the trending topics (or the ideas that tend to appear more often towards the end) to help you see what ideas you touched upon most recently as you were creating the content.
Help & Support
Contact us via this portal or open an issue on the plugin's GitHub page if you are having problems, have suggestions, or need assistance.
Your Data is Private
We do not store any data you process through this plugin — not even in the logs. Your Obsidian vault data stays private and is fully owned by you.
When you use the built-in AI, some of the statements (that are only limited to the graph structure you are processing) may be sent to OpenAI's API for processing, but their Terms of Service state that they don't use the data processed via API for training, so you should be safe. The advantage of using InfraNodus plugin to other tools is that you don't have to send the whole data. Rather, we generate a Dot Graph (compressed graph string) from your original text, so it adds another layer of privacy.
To learn more about our privacy policy, please, refer to the main document on www.infranodus.com/
Comments
0 comments
Please sign in to leave a comment.