Other Applications of PageRank
In class and on our homeworks, we spent a good bit of time discussing the topic of PageRank. However, PageRank is most often thought of in the context of how its creators, Google, would most often use it—global web searches. However, this idea of ranking search results based on outside sources that support a result’s relevancy has been and continues to work its way into new fields around the internet.
For example, as talked about in a number of other blog posts (and here: http://betanews.com/2015/06/19/google-now-on-tap-pagerank-for-apps/), Google hopes to change mobile app experience with Google Now by suggesting transitions to other apps without having to go back to the home screen, manually find a relevant app, and only then appropriately navigate to the desired information. Similar to how the Web shifted from a place where a user often manually typed in web addresses to one dominated by link connections, they hope to provide simpler and faster connection between apps. In order to do so, Google must establish a system that can make the relevant app suggestions for this feature to be useful.
But the applications of PageRank concepts can also be used on smaller scales than the two incredibly large domains of Web pages and mobile apps. For example, as detailed in the Margin Hound article, LinkedIn recognized an opportunity to take advantage of this concept of “voting” for relevant/reliable pages for recruiters doing talent searches. Even though they had an existing “recommendation” feature that allowed people to recommend colleagues or friends in more detail, they released and strongly promoted their skill “endorsement” feature. However, these endorsements are MUCH shorter—often just a word or two. For recruiters looking for people with certain capabilities, keyword search is vital. What distinguishes this from a regular keyword search, though, are the endorsements which act in a similar manner to web pages receiving “votes” from pages that link to them. After filtering certain results, they are left with a set of results containing people with certain skills that is sorted in order of how likely they are to be truly talented in a certain area.
After first learning about the concept of PageRank, I somewhat naturally thought of other applications in web environments (like the LinkedIn example), but did not as naturally realize the untapped potential of mobile apps. It creates an interesting problem (with enormous upside) for large companies like Google and Apple to solve: how to recreate the connectedness of the web in a mobile environment which is inherently not quite as conducive to connection between apps. With the release of iOS 9, I can see how Apple is trying to solve this problem with the “Back to previous app” button, but *in my opinion* this doesn’t feel at all optimal. I’m excited to see how app connectedness will continue to improve as technology does.
Links:
Linkedin Endorsements: How Might They Affect Linkedin’s Search Algorithm?