Technical Info

MostViral.News works by calculating a virality score for the popular headlines of the day among world news, not only retricted to the US, though news articles must be English for MostViral.News to work properly. We retrieve headlines in an automated fashion on a daily basis. To accomplish this, headlines are scanned daily from 6 AM - 6 PM CST. A virality score is then calculated for each of these headlines. The way we define "virality" is how fast an article's reach (audience) grows. Obviously, "likes", "shares", and "comments" all need to be taken into account, but weighted differently. A "like" represents the least amount of effort from a user, and thus indicates the lowest level of engagement. A "comment" on the other hand, most definitely indicates active engagement from the user. A "share" is weighted highest because it actually increases the potential reach/audience. Domain Authority is a metric (0-100) developed by Moz.com that describes how successful a domain is in search engine rankings. Higher search engine rankings indicate more popular sites as search engine rankings are based on the number of backlinks. Thus, to calculate the reach of a given article, its domain authority must also be taken into account. This leads to the following simple formula, where R is reach and A the relevant article:

R = A_{Domain_Authority} * (A_Likes + 1.25*A_Comments + 1.5*A_Shares)

It is then simple to formularize virality as shown below, where V = virality and t = time:

V = ΔR / t

This insures that all 3 possible interactions are weighted differently, along with the fact that domain authority is taken account. This guarantees that an article from an extremely popular news site will be weighted significantly higher than an individual's blog, for example. This does not mean, however, that it is impossible for an article from a lower trafficked site to become the most viral article of the day. This is because domain authority has a maximum value of 100. For example, sites like cnn.com (which has a domain authority of 94) receive far more than 10 times the traffic of those with a domain authority of 0-10. The factor by which a site's audience is taken into account thus has an upper bound. Thus, while a personal blog might only receive a few hundred daily hits, it is still possible (though not easy) for an article published theirein to overtake an article from cnn.com, even though the latter might receive millions of hits. If we simply used average daily traffic as a multiplier rather than domain authority, this would effectively become impossible. Additionally, this package is written in a modular fashion, so most any of the parts can be easily customized. The social networks that are currently searched are Facebook, Pinterest, and Reddit. The Stable Diffusion AI model is used to generate an image for the article. It is displayed alongside the article's official image.

Note that this is still a simplification of the problem. It does not take into account that actions from one user might contribute to the article's reach more than another. For example, sharing a link to a user that has a large number of "friends" or "followers" would increase the reach of an article more than sharing a link to a user with a relatively small number of "friends"/"followers". This will likely be addressed in a future version.