33 comments

[ 4.4 ms ] story [ 81.6 ms ] thread
Neat - I'll need to play with it some more. Any plans to auto-import data so you don't need to rely on people inputing 'what' , 'where' , 'why' ? eg. recent facebook/twitter/browsing history.

What's the stack look like?

Thanks :) I feel like it will only become useful if lots of people start using it. For example now local trends show up on the "worldwide" page, which won't happen anymore once there's enough input.

I'm thinking about how to automate this input process but currently I don't see an easy way of doing it. That info is not freely available anywhere.

Stack is Python/Flask/Javascript/JQuery/MySQL/SQLAlchemy/Bootstrap

Cool concept. Do you plan to ingest data from anywhere other than manual user input?
Def thinking about it but not sure how. Maybe Twitter API. Google trends does not have a public API I believe.

Other than that I'm planning to reward people based on their predictions of what could become a trend. The earlier they are and the bigger the trend the more points they get. Then votes from these people would weigh more in the algorithm which could potentially increase the accuracy and also incentivise people to input data :)

You might consider consuming the Reddit API to identify certain trends.
>Maybe Twitter API

Perhaps it has changed since I used it last, but Twitter's API will give you a very limited amount of data unless you're prepared to fork over the big bucks.

true, I completely forgot about that. Twitter is out then, maybe Reddit API as someone else suggested :) I'll look into it.
Let's talk, I might be able to help you out.
sure! email me at boogaerts.thomas@gmail.com
This is cool, I added some hot celebs because I am shallow.

It is quite open to abuse though and how can this be addressed I wonder? (There is a url in the celebs).

Thanks for sharing.

For now, manual moderation against the trolls, recaptcha against bots. I realise it's quite open to abuse right now.
A pity it has to be that way.
Just paywall it, or require a Facebook account to sign in first. That way you get the exact audience you want. ;)
Interesting, but what is the incentive to contribute?
Currently not much, that will be one of the challenges. Later on – gamification, how well are you at predicting upcoming trends in certain cities/categories.
Yeah, that would be my first thought as well. It seems like a challenge to attach users to something like this when it exists outside of an application they already use (eg. Facebook et al). I would be interested in an update down the line to see if you are able to capture regular usage and how you did it.
Would over-incentivising lead to more scatter-gun contribution and make the real trends hard to surface?
You'd probably have to limit the allowed number of "guesses" per day/week. But even then there's a risk of people trying to game the system (creating multiple accounts etc)
Nice straightforward UX. My first thought was to search for some trends about weather (everyone's talking about how hot it's going to be in southern california today) - but I didn't see that category.

So I'd suggest to consider including a section about weather/natural phenomena.

Seems interesting, nice interface. But why would I use this then google trends?
I like the idea but was expecting it to pull data from various sources like Twitter instead of just showing what people enter into it.
Same here, entered Germany and didn't get much results
Seems to already be filling up with jokes and memes.