Is there a website for tracking fads?

post by romeostevensit · 2019-12-06T04:48:51.297Z · LW · GW · 5 comments

This is a question post.

Contents

  Answers
    9 liam_hinzman
None
5 comments

Also, clever ways to use google trends etc?


Answers

answer by liam_hinzman · 2020-05-22T02:00:55.516Z · LW(p) · GW(p)

Treendly and Exploding Topics both show terms that are rapidly increasing in Google search volume.

A week ago I was curious to see if many people were cheating on the AP exams - they're online for the first time due to COVID-19. I searched "derivative" on Google Trends and saw that a ton of people were probably cheating due to a spike in search volume exactly when the AP calculus exam started (related searches also indicated what the exam question was).

comment by romeostevensit · 2020-05-22T04:33:42.827Z · LW(p) · GW(p)

These are pretty great!

5 comments

Comments sorted by top scores.

comment by Valentina V (valentina-v) · 2020-10-09T15:41:40.649Z · LW(p) · GW(p)

Free emerging trends detection tool https://emertrends.com/

comment by gwern · 2019-12-06T18:29:02.721Z · LW(p) · GW(p)

How would you define 'fad' in an objective and non-pejorative way?

Replies from: romeostevensit
comment by romeostevensit · 2019-12-07T00:27:33.620Z · LW(p) · GW(p)

Something currently undergoing the steep part of its sigmoid wrt memetic replication maybe.

Replies from: gwern
comment by gwern · 2019-12-09T23:27:16.802Z · LW(p) · GW(p)

Calling something a 'fad' has many of the same problems as calling something a 'bubble'. It's an invitation to selective reasoning. As Sumner likes to point out, most of the things which get called a 'bubble' never turn out to be that, it was just an insult and then a bunch of cherrypicked examples and flexible reasoning (think of all the people who called Bitcoin a bubble when it collapsed to a price far higher than when they called it a bubble).

I think you could get something more useful from a more neutral formulation, like specific cultural artifacts. So 2 recent relevant papers which come to mind would be https://www.nature.com/articles/s41467-019-09311-w http://barabasi.com/f/995.pdf . You could do a post hoc analysis and operationalize 'fad' as anything which rose and fell with a sufficiently steep average slope. (Obviously, anything which rises rapidly and then never decays, or only slowly decays, doesn't match what anyone would think of as a 'fad', or rises slowly and decays slowly etc.)

Replies from: romeostevensit
comment by romeostevensit · 2019-12-10T03:21:03.137Z · LW(p) · GW(p)

Ah, yeah model building from past fads would be useful. I am mostly interested in tracking things like current trends in a mostly content agnostic way, i.e. what is currently being shared the most across lots of platforms. Probably this is a paid service by some marketing firms.