Each topic on PopGraph is assigned a PopScore. This is a weighted average of all the scores previously calculated from other sources pertaining to that topic.

PopScores range from 0-10 and are rounded to 1 decimal place (e.g. 7.9). Higher scores indicate increased popularity.

The score is considered weighted because we assign more significance to certain scores from particular sources. For example, we may assign a greater significance to IMDB when scoring an actor such as Keanu Reeves than we would to, say, CNN.com. This is because IMDB, being a movie/actor related site is considered a higher quality authority on actors than CNN (which is about news and current affairs).

The final PopScore is also effected by the context of the topic and how that relates to other topics in the same context. For example, the band Metallica is not the same context as the musical album of the same name (Metallica), and therefore they cannot be applied with the same weighted scores and data inputs. In fact, context provides a very powerful system in which one can conceptually view a topic's popularity within context: for example how popular is Arnold Schwarzenegger simply as an "actor" versus how popular is he as a "politician", or as a "body builder"? The scores would likely differ because the context is different. Arnie is probably one of the most well known body builders ever, but not the most well known politician (think JFK, Obama..). However, context specific scores are a more advance feature, and the scores you see on the main site represent the overall popularity in the general context of "all things".

Question: So, how do you compute the scores?
Answer: As we said above they are computed using weighted averages, based on both the source and the context. In addition to this we use various types of data gathering and analysis techniques such as search result frequency, time trend analysis, score normalization, and much more. For more specific, focused and relevant results we also use the open APIs of various major sites and apply our own custom data gathering and score metrics to those on a individual basis. For example, the popularity of a person might be in part concluded from how many friends they have have on Facebook or how many followers on Twitter, right?

Question: What is "Score Pending"?
Answer: We can't track everything all the time. Some things we haven't scored yet. When that is the case it is pending, and usually you will see this if someone has viewed a topic within the last 24 hours that had no PopScore. We then know to go out and get a score for that topic and try to do so within 24 hours.

Question: Last week XXX has a score of 8.9 and now it has a score of 9.2. Huh?
Answer: Popularity changes. What is popular today may not be so popular tomorrow. We recalculate scores frequently to kep up to date. Think for example how popular Barack Obama was 10 years ago as compared to today!