future social media research
TRANSCRIPT
The Future of Social Media Research!Or how to re-invent social media monitoring in 10 steps Francesco D’Orazio @abc3d | PulsarPlatform.com
legacy!Since it emerged 15 years ago, the industry has been largely responsible for driving some of the most interesting evolutions in the research space, such as the democratization of text mining and computational approaches to mining qualitative information.
qual/quant!blurring!
This is an approach that for the first time enables both a granular and a birds-eye view of the data, making it possible to produce qualitative observations on a mass scale. A new perspective that is blurring the lines between qualitative and quantitative thinking.
a new epistemology !
And with computation also comes the ability to mine larger (and messier) datasets, which is in turn steadily shifting the focus of what we call knowledge, from understanding causation to identifying correlations.
Social media monitoring is a growing industry but one that is stuck in its old ways. And in need a of a urgent re-think. Or As McLuhan would have put it, we look at social data through the rear view mirror
the 480+ smm platforms are
broken!
Add to this the more systemic revolutions the industry is facing, such as the visualization of social media, which is going to pose huge challenges to an industry that’s been entirely built on text mining.
#tumblr: !85m post / day!84% images !
social media research is not web analytics!
It’s not quantitative
data!
It’s qualitative data on a
quantitative scale!
Drop the ‘media’. !
It’s ‘social media data’!
It needs social science not
‘media monitoring’!
How can social research
fix smm in 10 steps!
sampling beyond
keywords!
1!
09 am
10 am
11 am
10 pm 12 pm
1 pm
2 pm
3 pm 4 pm 5 pm
6 pm
7 pm
8 pm
9 pm
11 pm
07 May 2013, 10 pm: the rumour spreads on Twi?er
Great Britain Indonesia USA Malaysia
Spain India Ireland
France
Nigeria
Other
South Africa
How fast does news travel? Conversations about Alex Ferguson
retirement by the hour by country
4 maps by visibility
4 maps by visibility
@Brand follower activity!
Mentions of brand x!
0.1%!
Decoding your online audience Discovering clusters of fans using Social Network Analysis
"Detecting social communities in "Question Time’s Twitter audience
OwenJones84
PennyRed!
davidschneider! fleetstreetfox!
bbc5live!
bbcquestiontime BBCNews!
SalmaYaqoob!
johnprescott
mehdirhasan!
afneil!
ChrisBryantMP!
paulwaugh!
WillBlackWriter!
RippedOffBriton!
NHAparty!
richardcalhoun!
steveclarkuk!
oflynnexpress!
RicHolden!
Green Cluster - 28 y.o., White British - Students, Musicians and actors - From London (41%), Manchester (10%), Liverpool (4%) - Christian and Jewish - Into partying , reading, comedy, football; - Following @stephenfry, @RealDMitchell and @daraobriain
Blue Cluster - 35 y.o., White Birtish - Senior Managers , Journalists, Writers and Lawyers - From London (50%), Manchester (3%), Leeds (2%) - Christian and Jewish - Into Politics, dining and wining, tennis, football; - Following @David_Cameron, @stephenfry, @Number10gov
80%
20%
63%
37% 56%
44%
Fuchsia Cluster - 32 y.o., White British - Manufacturing professionals, Nurses, Teachers - From London (48%), Manchester (6%), Cardiff (3%) - Jewish, Christian - Into Business news, reading, history, rugby, tennis and golf; - Following @BBCBreaking @Number10gov @Lord_Sugar.
59% 41%
Orange Cluster - 42 y.o. White British - Manufacturing professionals, Nurses, Teachers - From London (48%), Manchester (6%), Liverpool (5%) - Jewish, Christian - Into politics, tech news, environment, cooking, tennis; - Following @BBCBreaking, @Ed_Miliband, @Queen_UK.
69%
31%
Pale Blue Cluster - 20 y.o, White British, Black - Manufacturing professionals, Nurses, Teachers - From London (48%), Manchester (8%), Belfast (4%) - Muslim, Christian - Into Gaming , comedy/humor, sports; - Following @jimmycarr, @andy_murray, @rioferdy5
2! from content to
context!
© Alexandre Farto aka Vhils 2010
Most Social Media Monitoring platforms focus on just Content
© Alexandre Farto aka Vhils 2010
But we also wanted to understand Context and Behaviour
Social Graph Interest Graph
Content Demographics Behaviours
Dimensions of Social Data
Nose-to-tail indexing or Bigger Big Data
Content
Behaviours
Social Graph
Interest Graph
Demographics
Data anthropology
3! from analytics to intelligence frameworks!
Analytics Intelligence How many
Tweets/Hour?
How many negatives?
What time of the week is best for what? What’s the brand equity?
Measuring Visibility
4!break down
the social silo!
Predicting the Oscars Integrating social data with multiple external datasets to increase predictability
EXP #1 Reality Mining
Images
Location
Call Log
SMS Log
Bluetooth
Accelerometer
Gyroscope
Orientation
Activity
Running Apps
Battery
Screen
Browser Searches
THE PICTURES AS TEMPORAL AND SPATIAL MARKERS OF THE JOURNEY
5!scalable
human-analysis!
6!machine-learning!
7!dataUX !
A data canvas
Visual Mining
Data Transparency
8!decision-making!
9!hybrid
methods!
Real-time Audience Insights
Dynamic Segments
SOCIAL PANELS > real-time segments
Pulsar | Social Data Intelligence @pulsar_social | PulsarPla]orm.com
50k readers
23,000 Tweets
50k readers 50k readers
50k readers 50k readers 50k readers
22,000 Tweets 47,000 Tweets
63,000 Tweets 32,000 Tweets 12,000 Tweets
A Nation Divided? A Social Panels case study
The Daily Mail 23,000 Tweets
The Guardian 22,000 Tweets
The Independent 47,000 Tweets
The Daily Mirror 63,000 Tweets
The Telegraph 32,000 Tweets
16%
23% 8%
26%
10%
19%
8%
23%
15%
18%
12%
19%
Positive towards Thatcher
Negative towards Thatcher
Neutral: hidden
A Nation Divided? How 300,000 readers of six top UK newspapers are feeling about the death of Margaret Thatcher
10!
making research
programmable !
Social Data Intelligence
@pulsar_social | PulsarPla]orm.com