Wikipedia is one of the top 10 most widely used websites globally and a rich source of information. There are pages for every conceivable topic; the English version alone has over 5 million. Using Wikipedia it is both possible to rapidly get information about a topic and then to also delve deeper into the details.
However, often it can also be useful to quickly see other pages that are related to the one you are reading. This provides broader context or a more lateral view of a topic and also means that you can quickly find the most relevant other pages to read, without having to sift through all the links on a page.
With that in mind we created the Chrome Extension Similar Pages for Wikipedia - a simple extension to Wikipedia that displays similar pages at the top of every Wikipedia page:
So far a lot of feedback we have gotten has centred around how our recommendations compare to those offered by Wikipedia's related pages beta feature and building an integration for Wikiwand - a popular Wikipedia reader.
Wikiwand actually also offers related content at the bottom of Wikipedia pages, so below is an example of how the three compare:
and here's another:
As you can see all three perform very well. We currently use only the content of the Wikipedia page to show similar pages, but will begin using our hybrid recommender to incorporate user behaviour, which is likely what Wikipedia is currently using to power their related content recommendations.
What differentiates what we currently offer from the other two is that we place the recommendations at the top of every Wikipedia page and that you can choose to see up to 10 similar pages.
Our technology can also be used to provide similar Wikipedia pages for any webpage or text on a webpage. Using comparable functionality to our NewsBot Chrome extension, v2 could display similar Wikipedia pages as desktop notifications:
Additionally, we can also provide the option of displaying similar news or academic articles at the top of every Wikipedia page, as an alternative to showing similar Wikipedia pages.
Please let us know of any features you would like to see in v2 and we hope you enjoy the extension!
Demonstrating how to generate a dataset for recommending Reddit posts based on semantic similarity.
Today we want to show how you can make a spreadsheet based personal research assistant in just a few minutes using Blockspring and Lateral.