InfraNodus can be used to analyze the data from CrunchBase to identify the top investors in your industry, to reveal the key topics and keywords used by the companies when they talk about their product, as well as the adjacent industries.
1. Export the Data from CrunchBase
In order to export the data, you will need CrunchBase Pro account and download the search results using their CSV export function.
In our example, we are making a search for all the companies that work in "text analysis" and "natural language processing" industry (the total of 113 results).
Make sure to add the columns: "Industries", "Description", and "Top 5 Investors" into the export view as we might need this information later for analysis:
You can also download the data yourself using a browser plugin like Artoo. Note, that we do not encourage or recommend scraping, so, please, perform this only if you have a CrunchBase Pro account or have a legal right to do that. Having said that, we don't quite understand CrunchBase's policy to not allow CSV export on trial accounts, so here is a little snippet you can use to export and scrape the data using Artoo on Crunchbase Pro account:
let csvResultsPage1 = artoo.scrape('grid-row', {
name: function() {
const companyName = $(this).first('grid-cell').find('.identifier-label').text()
if (companyName) return companyName
},
industries: function() {
const industriesName = $(this).find('.column-id-categories').text()
if (industriesName) return industriesName
},
description: function() {
const description = $(this).find('.column-id-short_description').text()
if (description) return description
},
investors: function() {
const investors = $(this).find('.column-id-investor_identifiers').text()
if (investors) return investors
}
});
// repeat the above for every page
// combine the results
let csvResults = [...csvResultsPage1, ...csvResultsPage2, ...csvResultsPage3]
// export the results as a CSV file
artoo.saveCsv(csvResults, {filename: 'crunchbase_results', delimiter: ';'});
2. Import the Data to InfraNodus
Now that you have the data, you can import them into InfraNodus using the CSV import app.
You need to choose the column you will be analyzing. We recommend that you select "Descriptions" first. This will allow you to see what language the companies that work in this industry are using when they talk about themselves. This may be useful to better understand the market and to extract the buzzwords used in the industry:
3. Get an Overview of the Industry's Discourse
After a few seconds, you will see a text network graph of the companies' descriptions:
Using the graph and the Analytics panel on the right, you can see the main keywords used are:
"text", "ai", "data", "analytics".
The main topics are:
"data", "unstructured", "insight"
"ai", "customer", "platform"
"text", "analytics", "search"
"machine", "human", "empower"
As we can see, some of the key buzzwords in the industry of text analysis and natural language processing (apart from the main terms like "text analytics") include "unstructured data insight", "insight", "ai", "search" and "empower".
We can make a hypothesis here that most of the companies listed on CrunchBase that work in the field of text analytics and NLP provide insight into unstructured data and improve search. They also talk about "ai" and "machine learning" but these are not the main buzzwords in this particular discourse. There is also an emphasis on platforms and customer data analysis.
4. Explore the Specific Concepts and Relations Between Them
Let's test our hypothesis by clicking the nodes "machine" and "learning" on the graph to explore what other concepts they are both connected to:
Click on the Analytics > Relations panel to see the top relations present in this discourse:
"text", "ai", "data analytics", "platform" and "intelligence"
On the left-hand side, you can also see in which contexts those terms are used and how many percent of company descriptions use the term "machine learning" (8%) — not so many.
We can say that in this particular industry, the notion of providing insight into unstructured data and concrete use cases (e.g. customer feedback analysis, customer support) are more important than the actual technical solution used (e.g. machine learning or AI) although it still plays an important role.
5. Explore the Nuance — Removing the Most Frequent Terms
The next step is to remove the most frequently used terms in order to see what's hiding behind them and explore the nuance in this discourse. You can select them on the graph and then click the "hide" button or use the "Reveal Non-Obvious" button in the Analytics > Topics panel:
As we can see, the main topics that come up, as a result, are the
"platform", "unstructured" [data] "analysis" (confirming our previous hypothesis)
and
"customer", "insight" and "voice" (the most common use case: customer feedback analysis)
We also see a cluster that contains the keywords such as "search", "develop" and "ai" — let's click on the node "search" to explore in which context it appears:
As we can see, it is used in several product descriptions, but the context of where it's used is quite diverse.
6. Summarize Your Findings
Now that we discovered the main topical clusters in CrunchBase company descriptions, we can summarize our findings using the built-in AI.
Go to the AI Insight Panel and click on "Edit the AI Query" button. Then click "Summary" on the Analytics > Topics panel or enter our hypotheses manually:
"text", "ai", "data", "analytics" and "data", "unstructured", "insight", "ai", "customer", "platform", "text", "analytics", "search", and "machine", "human", "empower", take into account "customer voice".
After generating a few results, we will get:
A text analytics platform that uses machine learning to empower customers with insights from their data. The platform helps companies search and analyze their unstructured data (e.g. customer reviews) for insights that can be used to improve their business.
or
Our platform helps you gain actionable insights into unstructured text data. Using our AI, you can analyze any discourse — from customer product reviews to research papers — and reveal the main topics and gaps between them.
7. Reveal the Structural Gap = New Business Ideas
You can also use InfraNodus to reveal the gaps in the current discourse: the topics that could be better connected. Our research shows that those gaps are a source of new innovative ideas as they connect the topics that are already present within a discourse but are not yet connected.
Go to the Analytics > Gap Insight panel. Then click Reveal the Gap button:
As you can see, one of the propositions is to think of an application that improves the quality of online content.
It takes into account the topics that are already present in this industry's discourse, and proposes a question that would connect them in a new way.
In our case, we can modify our initial summary with the following text:
A data analysis platform that uses network analysis and AI to gain insight from unstructured text data. It can be used for better understanding any existing discourse (e.g. customer feedback, research papers, online conversations) in order to create quality content that will touch upon the important topics connecting them in innovative ways.
Comments
0 comments
Please sign in to leave a comment.