8. AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
AND THEN SOME MORE DATA
YET EVEN
MORE DATA
and so on,
and so on,
and so on…
DATAAND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
AND THEN
SOME MORE
DATA
EVEN
MORE
DATA
14. 87%
Guest ID assigned
to every customer
Tied to credit card,
name, email
address.
Purchases analyzed
and commonalities
from pregnancies
identified.
Combine for score that predicts both
pregnancy and due date…
18. 1. IDENTIFY A QUESTION TO BE ANSWERED,
A PROBLEM TO BE SOLVED
2. COLLECT RELEVANT INFORMATION ON THE ISSUE
3. ANALYZE THAT INFORMATION
4. FORM A CONCLUSION
Effective use of data
19. “But @theelusivefish,” you tweet, “that’s nice
and all, but what does that have to do with
social media data?”
“YEAH!,” tweet others in agreement and then
RT, “that’s nice and all, but what does that have
to do with social media data?”
“hashtag hashtag hello, I’ve got a great deal on
viagara cialis bit.ly/343q2a” tweets an
anonymous bot that saw activity on the topic
and dived in for the sale.
Because it’s not about ‘social media’ data. It’s
about ‘DATA’. Like people, you want to treat
data the same no matter where it came from.
20. But let’s talk about a familiar
social media tactic:
‘Online Influencers’
22. Who would have ever thought that one
day, I would be an influencer on the
information super highway? Come
closer and touch my sweater… some of
the klout may rub off on you.
Does it work, or are we just
giving away free stuff to
popular people?
23. $$
$ $
25 boxes of swag
20 tours of factory
18 warm handshakes
from the CEO
150 tweets
6 Facebook posts
4 Blog posts
4 RTs
2 Favourites
10 Likes
Normally
we report…
But we want to know
outcomes not outputs
25. Compare all
brand mentions
by our audience
with brand
mentions by
those not in our
audience.
THE METHOD:
Audience mapped using NodeXL
Raw data obtained using Radian6
Analysis conducted in Excel
29. What’s changed?
ANYTHING CAN BE MEASURED.
ANYTHING.
LOOK FOR WHAT CHANGE YOU ARE SEEKING TO MAKE.
If it changes, then it can be
measured.
30. “Data! Data! Data!”,
he cried impatiently,
“I cannot make
bricks without clay!”
- The adventures of Sherlcok Holmes
31. Identifying
contextual
relevance from
peers and
volunteered info.
Turtles
Shelled Animals
Turtle Owners
Turtle Turtle Turtle
Reptiles – Snakes and Turtles
Tortises
Turtle Lovers
Turtle List
Subject
experts on
Turtles?
Extract lists via NodeXL
Analyze in Excel
32. Don’t just
leave it to the math
The social gestures don’t always
mean what we think they mean.
This guy likes turtles.
This guy …
not so much
33. Please don’t
pee in the pool
Encouraging false
signals spoils the
pool of data for
everyone
34. 80% of data analysis
is cleaning the data
Keep your data TIDY!!!
OpenRefine as a useful tool for cleaning messy data
openrefine.org
35. Structure your data
FIRST VARIABLE SECOND VARIABLE
1st Observation Value Value
2nd Observation Value Value
3rd Observation value Value
One type of observation
• Each table contains one type of observation
• Each column is a variable
• Each row a set of observations
Keep your data TIDY!!!
36. 1. IDENTIFY A QUESTION TO BE ANSWERED, A PROBLEM TO BE SOLVED
2. COLLECT RELEVANT INFORMATION ON THE ISSUE
3. ANALYZE THAT INFORMATION
4. FORM A CONCLUSION
Use your data effectively
Look for the change or predictable patterns
1. COMPARE OVER TIME
2. COMPARE TO SIMILAR ‘CONTROL’ GROUP
3. USE GROUP WITH KNOWN OUTCOMES TO FORM PREDICTIVE MODEL
Keep your data tidy
1. COMPARE APPLES TO APPLES
2. STANDARDIZE AND STRUCTURE OBSERVATIONS
38. CREATIVE COMMONS – BY ATTRIBUTION
Data action figures
http://www.flickr.com/photos/jdhancock/8031897271/
LinkedIn Dataset
http://www.flickr.com/photos/luc/5418037955/
LEGO Wizard
http://www.flickr.com/photos/tamaleaver/7419788172/
Gnomes
http://www.flickr.com/photos/debabratad/4569271310/
Target shopping cart
http://www.flickr.com/photos/jreed/1352409015
Pregnant belly
http://www.flickr.com/photos/nateone/3240716239/
Purse
http://www.flickr.com/photos/lululemonathletica/4209254375
Cocoa Butter
http://www.flickr.com/photos/franciscouhlfelder/5404451647
Blue Rug
http://www.flickr.com/photos/andie712b/4619093237
Target right behind you
http://www.flickr.com/photos/gazeronly/6150401295/
Ford Online Influencers
http://www.flickr.com/photos/pandemia/6731863193/
No public access
http://www.flickr.com/photos/jmv/2734200159
Benedict Cumberbatch on set
http://www.flickr.com/photos/bellaphon/4409531705/
Pee in the pool
http://www.flickr.com/photos/lifeontheedge/230245129
Photographer
http://www.flickr.com/photos/evilerin/3098610791
Photo Credits
Public Access photo by jmv - http://www.flickr.com/photos/jmv/2734200159Benedict Cumberbatch on set for Sherlock by bellaphonhttp://www.flickr.com/photos/bellaphon/4409531705/
Turtle - “Jonathan Zander (Digon3)“http://en.wikipedia.org/wiki/File:Florida_Box_Turtle_Digon3_re-edited.jpgKrang - http://mikev.me/ Mike V
Photo by Marshall Astor – Food Fetishist http://www.flickr.com/photos/lifeontheedge/230245129
Photo by Tama Leaverhttp://www.flickr.com/photos/tamaleaver/7419788172/
Photo by Evil Erin http://www.flickr.com/photos/evilerin/3098610791