Topics and Categories are new tabs in the Historical Dashboard to explore content performance with the help of AI. These new views of content performance allow publishers to understand larger patterns of what kind of topics and coverage are driving engagement so they can make strategic decisions about what to dig into deeper.
- What are Topics?
- How does Chartbeat determine an article’s Category?
- Which languages are supported by these features?
- Is there a plan to add more languages?
- Can the system pull data from paywalled articles?
- What IP addresses does Chartbeat use to visit pages?
What are Topics?
Topics are AI-extracted entities from the text of the article. These may be persons, organizations, geopolitical entities, events, works of art, or products. We use AI to determine which of these entities are important to the text of the article, and can be considered part of the topic of the article. The number of topics assigned to a single article varies depending on the length and content but may typically be in the range of 3 to 15.
How does Chartbeat determine an article’s Category?
Metadata is extracted through the text through natural language processing techniques. We then assign up to three categories to an article based on its content using IPTC taxonomy:
- Religion and belief
- Arts, culture, entertainment and media
- Disaster, accident and emergency incident
- Crime, law and justice
- Economy, business and finance
- Science and technology
- Conflict, war and peace
- Lifestyle and leisure
- Human interest
Which languages are supported by these features?
This feature is currently available to all sites that publish in English, Spanish, and Portuguese.
Is there a plan to add more languages?
We plan to gather customer feedback from clients via an Early Access Program. We will consider expanding the languages we support.
Can the system pull data from paywalled articles?
Non-paywalled articles are able to be scanned for topics and categories today, but Chartbeat may fail to correctly identify the text of the page due to paywalls or popups which block the content of the page.
You can increase our ability to detect topics and categories on your pages by allowing traffic from our IP addresses (listed below) to bypass any features which obfuscate the content of the page. If you are not able to do this or have any questions, please contact us at email@example.com.
What IP addresses does Chartbeat use to visit pages?
Chartbeat regularly tries to visit pages where we detect significant traffic to examine the content of the page. You may notice traffic on your page from Chartbeat (note that this traffic will not appear in your Chartbeat dashboards), and you can identify our visits as they will be from the IP addresses listed below:
Please note that while this list does not regularly change, it may be expanded in the future due to infrastructure improvements.