Live data from Hacker News

Show HN: DeepSense – Know people better through AI-built profiles

frrole.ai

41–48 of 48 posts

Re: Show HN: DeepSense – Know people better through AI-built profiles

#41
post #28

This is actually a really hard problem to solve. At my company we have been working with several social data enrichment providers and even the "best" have a pretty high degree of error. For what it's worth we just typed in a few "lesser-known" emails/twitter handles and it picked up relevant and accurate data. I'm going to signup for an API key and fill out the contact form so we can talk to you guys. Is there any pl…

Thanks Ryan, this is not fully baked yet but def. not half-baked either. It will do what it is supposed to do. Instagram is coming, and linkedin too. Soon, very!

Totally agree with Ryan, I signed up API email right away too. This is something dope @kalkat,great work!

Re: Show HN: DeepSense – Know people better through AI-built profiles

#42
post #37

Earlier quoted context omitted.

Following is based on using his Twitter profile as an identifier: "Donald J. Trump is a spontaneous individual and takes life as it comes. he generally dislikes coloring within the lines and does not stress over the little things but sometimes struggles with procrastination. He is a very dynamic individual who is easily excitable and sensitive. Passionate and impulsive, he wears his emotions on his sleeve. He is slig…

I wonder if it filters out retweets in the analysis. Would be interesting to see the differential. Also the differential for media included (video and photos of thank you messages) vs the direct raw 6am-10am messages.

yes it does. More about filters in the api docs here: http://api.frrole.com/v4/docs/user-profile

(warning: docs not fully updated yet. usual development to documentation lag:))

Re: Show HN: DeepSense – Know people better through AI-built profiles

#43

Is this sort of analysis technically allowed by Twitter's ToS and Privacy Policy? Don't get me wrong, what you post on social media is public and that's that, but I know there are sometimes interesting rules around this sort of analysis.

"this sort of analysis" is a bit wide ranging, so can't comment, because some 'such' analysis might not sit pretty with twitter. But "this" is in line with policies, we in fact work with twitter and twitter is aware of this application.

This in fact has the potential to make twitter data useful for a new kind of user (product manager at a software product), so has the potential of making twitter useful for a sort of new segment of users.

Re: Show HN: DeepSense – Know people better through AI-built profiles

#45
post #28

Earlier quoted context omitted.

Thanks Ryan, this is not fully baked yet but def. not half-baked either. It will do what it is supposed to do. Instagram is coming, and linkedin too. Soon, very!

Totally agree with Ryan, I signed up API email right away too. This is something dope @kalkat,great work!

thanks mate!

Re: Show HN: DeepSense – Know people better through AI-built profiles

#46
post #25

What is with this user activity plot? http://imgur.com/a/3dgCp

It's the volume of the user's posts on social media over the months.

But why is it plotted with this arc bars. That is more than weird, isn't it?

Re: Show HN: DeepSense – Know people better through AI-built profiles

#47
Based on my Twitter account, I learned that I'm "extremely negative and pessimistic", while "a generally positive individual". I'm also "even tempered" even though "emotional and judgmental".

(Full generated profile below) "[my Twitter user name] is quite compassionate, cooperative, even tempered and helpful and is sometimes taken advantage of because of his trusting nature. He is extremely negative and pessimistic. He is emotional and judgmental. He is a generally positive individual and likes being around other people. He has a good social circle and can be assertive when needed."

Re: Show HN: DeepSense – Know people better through AI-built profiles

#48
post #19

From the responses here, it seems that astrology provides more useful insights.

Especially given the use of generic, meaningless phrases such as "is sometimes taken advantage of because of his trusting nature". Generated profile can even be self-contradictory (I posted mine in another comment - it said both that I'm "extremely negative and pessimistic" as well as "a generally positive individual"). Someone is jumping the AI bandwagon...
Post reply on HN