Amazon Kindle readers use highlights to save book quotes that they like. While it’s possible to later review those highlights using the Amazon’s own readers, including the online one, as well as third-party tools like Readwise, the approach is based on memorization rather than generation of new ideas.

It seems like ideation is one of the purposes of reading and so it is strange not to have the tools that help generate ideas from books. In this article, we introduce a system that is based on text network analysis of Amazon Kindle highlights to reveal the main topics and ideas in them. Using advanced graph analysis measures we then show how the knowledge graph can be used to identify the gaps between ideas and generate interesting research questions that would bridge these gaps in new ways. A good idea is often in relation to something that already exists, but making a new connection, taking a different direction or even going outside of the periphery to bring ideas from a completely different context.

Using the text network analysis tool InfraNodus, we generate a knowledge graph of Amazon Kindle highlights using the InfraNodus browser extension. Following the tutorial “Readwise Alternative: Knowledge Graph“, we reveal the main clusters of ideas (topics) and explore the relations on the graph. This gives us a good overview of the main ideas in a particular book.

For example, if we analyze the highlights for the book “The Geography of Thought: How Asians and Westerners Think Differently – and Why” by Richard E. Nisbett, we will get a direct visual overview of the main ideas inside:

As a result, we can see that the book’s highlights are talking about Japanese, Chinese, and Western cultures, their contradictions and constraints, as well as beliefs and thinking processes. It’s exploring how these idiosyncrasies emerge in human attributes, within different cultures, and through the events of cultural significance. Note, that we can either make an effort to link those ideas together or use the built-in AI “summarize” function to do the same. However, performing the exercise of connecting the ideas highlighted on the graph into a coherent statement is a very good practice for internalizing the meaning of the book’s highlighted ideas.

With the help of AI, we can see that specifically there is a focus on the concepts of individuality vs. collectivism, the role of context in perception, and the differing approaches to logic and harmony.

After formulating the main idea of the book, we can go further and look for the gaps between those ideas. InfraNodus builds its knowledge graph based on co-occurrences of the concepts used inside the text. Those ideas are organized into clusters to reveal the main topics they form. There will always be blind spots between these topics: groups of ideas that are not so well connected or that are underdeveloped. These are the structural gaps that InfraNodus will propose to bridge with interesting research questions.

Again, it may be more beneficial for ideation and learning to make a conscious effort to think of possible questions that could bridge these topics to activate a creative thinking process, following the ecological variability framework we presented before. It is also possible to use the built-in AI to generate those questions and answers:

As a result, the reader is motivated to think how to develop connections between the topics and ideas presented in the book further. These research questions can be regenerated multiple times on multiple structural gaps identified in the graph.

This methodology can also be used without any digital tools. For any book that is being read, there is always a web of concepts that emerge during the process of reading. These concepts can be organized into topical clusters to get a general (zoom-out) overview perspective of the text. We then zoom in and attempt to connect those topics in a new way by asking ourselves: what if we link this idea that the author is talking about in this book to that other idea that was mentioned at another time but wasn’t connected to the first one? This will encourage ideation process and coming up with new ideas based on the book’s content.

The next step is to compare the highlights and quotes to the content of the book to see whether something is missing. This can also be a really good way to check one’s own cognitive bias in order to see if only a certain aspect of the book was highlighted and whether there are topics omitted. We will demonstrate it in another case study that will be published soon.