How to Import JSON Files Content and Visualize Them As Knowledge Graphs

InfraNodus can be used to import the JSON file content and visualize them as knowledge graphs. This will help you visualize the patterns present in the JSON file (in the fields you select) and you can also use category fields to add content filters.

Let us demonstrate how it works step by step:

1 Open the JSON Import App

All you need to do is to open the JSON import app and upload the JSON file you want to analyze. For instance, we can take the file from the Financial News JSON dataset from Kaggle which contains a table of news headlines and associated events. 

2 Select the Fields to Analyze

In the next step, select the JSON fields you want to run text analysis on. InfraNodus will automatically analyze every field with this name if there are multiple fields in the file:

In our case, we're choosing to analyze the news headlines.

3 Select the Fields to Use As Filters

The next step propose to select the field values you'll use as filters. It makes sense to select the fields that have a lower variety than the number of values in the file. Usually — 8-15 categories you can use to filter this data:

In our case, we're choosing the type of event as the filter.

4 Additional Settings

If you don't want the phrases in a certain JSON field you're analyzing to be split into concepts or entities, you can choose to analyze the whole value in the field. In that case, InfraNodus will add [[squarebrackets]] around the value in the field in you're analyzing, so the whole value is treated as entity. This makes sense if you're analyzing several fields and want to see connections between the values in the fields rather than analysis of the content inside those fields. 

5 Generate a Knowledge Graph from JSON

Now we can generate a knowledge graph from the JSON file:

This graph shows us the main topical clusters identified in the news titles:

• Market surge
• Positive sentiment
• Tech rally
• Trade signals

We can also use the graph to identify the gaps between the topics: for instance, between the trade signals and tech rally, and use the built-in AI to generate research questions and ideas to bridge them. 

6 Use Filters to Get In-Depth Insights

You can also use the filters we added in step 4 to filter the topics by a certain event type. For instance, when we apply the filter by the event "Cryptocurrency Regulations" we can see that there's an emergence of the "consumer confidence and spending" cluster, which means there's an association between this type of event and this type of the news cycle: