🎥 Filter the Graph and Statements by Tags / Categories: Metadata Tagging & Segmented Analysis

Using InfraNodus, you can apply metadata tags to each statement added to the graph. This allows for category-aware filtering, so you can explore and analyze the graph by specific dimensions or segments that matter to you.

This approach enables segmented analysis of text data based on defined attributes. For example, if you import daily news from five different publications, you can filter by media outlet to examine the distinct topics each one emphasizes. Similarly, if you import survey responses in a CSV file where one of the columns represents a demographic variable (e.g. gender, location, or age bracket), you can apply that column as a filterable dimension and analyze the responses by those specific segments.

Video Tutorial:

1. Adding Metadata Tags to Your Data

There are several ways to do metadata tagging in InfraNodus. The general idea is to tag the statements by tags which you can then use to filter the statements and highlight / filter the graph.

There are 3 different ways to add metadata tags to your text statements:

1. Automatic tagging using InfraNodus

2. Tagging data using CSV files (via columns that tag the rows, which are treated as statements)

3. Manual tagging using the text editor inside the graph 

To learn more about how to add tags to your data, please, check the help center article How to Add Metadata Tags and Attributes to your Graph Segments and Statements

2. Filtering Graphs by Tags

Let's suppose you created a graph from the following spreadsheet showing the search queries used to find InfraNodus, the country they originate from, type of intent, and their popularity:

As a CSV file, it will look something like this (you can download it here):

"Search query","Country","Type","Popularity"
text analysis tools,US,Academic,70
thematic analysis,UK,Academic,60
network analysis,UK,Business,50
data visualization,US,Business,55

When you import this file as a CSV, InfraNodus will ask you specify which columns to use for filtering, you can choose "Country", "Type", and "Popularity". 

The text data in the "Search query" column will be used for generating the graph. Each row is a statement with the value from the "Search query" column.

As a result, you will see a graph visualizing the search term co-occurrences that reveal the main topics and keyword combinations in search intent:

The filters that are available to you are shown with the statements above (the blue chips under each text). They are also available to you now in the filtering panel:

You can select some of those filters to show the statements that contain those filters in them. By default, InfraNodus uses the "OR" parameter, which means you will see the statements that contain ANY of the tags you select.

For instance, if you select "Country: US" and "Type: Academic" in the filter above, you will see 3 out of 4 statements (75%), which contain the "Country: US" or the "Type: Academic" tag:

This is why "network analysis" row from the original CSV file is not there: because its country category is "UK" and its type is "Business". 

3. Finding Segment Overlap: Showing Category Tags Intersections

In the example above we show a combination of different tags. However, in some scenarios, you might want to see how those categories overlap. For example, you might be interested to see the search queries that are only from the "Country: US" AND at the same time have the "Academic" category.

In order to do that, you apply the "Overlap" filter in the Filter panel:

As a result, InfraNodus will filter all statements to show only the one that has BOTH of the tags you selected at the same time ("AND" operator) and highlight this statement in the graph:

You can now clearly see that if you want to focus on the search queries from US that relate to the academic context, you should focus on "text analysis tools". 

3. Recalculate Analytics and Filter the Graph by Categories

You can apply additional filter, which will remove the parts of the graph related to the hidden statements, so you only see the concepts and AI-generated topics relating to the statements you filtered out. 

In order to do that, open the Filter panel and select the "Filter & Recalculate" option below:

As a result, your graph will only visualize the nodes that relate to the statement that contains both of those tags (as we're in the overlap mode):

Notice, how you get a much more precise topical definition here in the Analytics panel on the right.

In the "Highlighted" mode (section 2) you had more general topics ("analytical tools" and "visual data") because you still had the context of the whole graph.

In the "Filtered" mode (this section) you have a much more precise topical definition ("text insight") which relates to the content of the statements you filtered (and not to the rest).

This is particularly useful when analyzing large data sets. 

You can explore more complex filtering workflows in the tutorial below.

Related Tutorials

Learn how to use tag highlights and filters for practical use cases related to marketing and SEO:

- How to Use the Google Search Console Keyword Research Tool for SEO

- YouTube SEO Tutorial: Find Low Competition Keywords with VidIQ Keyword Research Tool

- Visualize Your Ahrefs Content Gaps: Competitor Research Guide 

- Visual backlink anchor text analysis using Ahrefs and AI-powered knowledge graph