Twitter 2.0 = Curated Signals + Applied Intelligence + Stratified Inference


P.S: Copy of my blog in Linkedin

Exec Summary:

One possible trajectory and locus (“product cadence”) for Twitter 2.0 is to be a platform – to tell stories with different levels of abstraction – from basic curated signals to aggregated intelligence (ie trends, positions, sentiments and issues) & finally the higher order of exposing stratified inference built on the signals and intelligence.

For example CPC advertisers might want to know “Who is an NBA Fan” for personalized ad campaigns based on the interest graph (We did a similar project few years ago, based on Twitter data)

All without sacrificing the core Twitter consumption experience, but adding different dimensions to Twitter consumption …

Constituents like political campaigns (pardon the pun ;o)) can consume the platform at different levels and sophistication. All the (potential!) President’s Data Scientists can run multiple models over the signals while All the President’s Devops can build dashboards for the strategists to consume the curated inferences

Twitter Network != Facebook Network; Twitter Graph != LinkedIn graph ie. Twitter is an interest graph, not a social graph. If so, why can’t Twitter expose the interest graph as a first class entity, with appropriate intelligence?

Twitter is the right platform for ad-hoc,ephemeral spaces to exchange quick notes.

[Update : Julia Boorstin’s blog What’s Next for Twitter also echoes many of my recommendations below]

Details:

Due to various reasons I have been contemplating about Twitter in general and specifically Political Campaigns as an example of an eco system where Twitter has lots of potential

Twitter has been in the news recently with the CEO change as well as the stock dip. Time for Twitter 2.0 ? Definitely !

Interestingly I had written about Twitter 2.0 in 2011 and most of it is still true ! I will include relevant parts from that blog

For the technically minded who are into the gory details, pl refer to materials from my 2012 OSCON tutorial [Social Network Analysis with Twitter] 

What do campaigns want ?

They want curated inference (which they can directly consume for actionable outcomes) and curated intelligence (for overlaying specialized models over the exposed signals at different orders). All the President’s Data Scientists would have interesting data science models over the Twitter signals. A general model is like so:

Twitter 2.0 – Trajectory & Locus

Now Twitter is an agora for pure message-based interactions; but it has lot more potential – to be a platform (of course,without sacrificing the essential nature of the medium) ! To get there, it needs to be proactive, providing different levels of abstraction – from the basic curated signals to aggregated intelligence (ie trends, positions, sentiments and issues) and the higher level of stratified inference. It also should provide congruences on Twitter to Rest-Of-The-World ie how indicative are the twitter-verse of the general population.

Topic Streams a.k.a TweeTopics


I use Twitter for 3 things – to keep current with topics that interest me, keep in touch with friends & acquaintances and finally publish things that I am interested in – many times as a bookmark !

It is almost impossible to follow topics. The List functionality never worked for me. It should be as easy to follow & unfollow at the level of topics. In the day and age, it is not that hard to run the tweets through a set of analytics engines, cluster them by subjects and offer the topics, with the same semantics as people ! The current interaction semantics are very relevant – that is what makes Twitter Twitter.

There was some thoughts about tweet threading – I think that defeats the purpose; tweets are stateless and that attribute is very important

Twitter is different from facebook and linkedin, it is not a social graph but an interest graph. Many of the traditional network mechanisms & mechanics, like network diameter & degrees of separation, might not make sense. But, others like Cliques and Bipartite Graphs do

Why can’t Twitter expose the interest graph, with appropriate intelligence ?

Topic Spaces a.k.a. TweetSpaces a.k.a TweetRooms

Twitter is the right platform for ad-hoc,ephemeral spaces to exchange quick notes.

This was my observation in 2011, and still it is true.

IM is too heavy weight and not that easy for quick things like “Where is that meeting room” or “Which seat are you in” or “What should we discuss next” et al. A one-to-many exchange, between people who are spatially (and even temporally) in separate spaces. They might in a plane, on a call or even in a hallway! Should be easy to add  a “!” tag, and shout the info. Yep, folks need to know what the ! tag is. Actually come to think of it, we could have many types of tags using a lot of the ‘$’,’@’,’%’,’^’,’&’ and ‘*’ characters with different semantics!

Time for a “Tweet Mark-up Language” ?

These are some of my quick thoughts, what says thee ?

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s