Big Data in Pharma USA 2014.

1.
Transforming
Informa-on
to
Liquid
and
then
Channeling
Liquid
Insights
out
to
the
Right
People
BigDip
Boston
2014
Sebas3en
Lefebvre
Director
R&D
IT
Pla?orms
BiogenIdec

2. Problem
statement
Enterprise
Content
Management
restrict
access
to
most
people
Increasing
partnerships
and
collabora3ons
across
the
industry
Increasing
internal
and
external
complexity
of
informa3on
sources
Provide
simplified
access
to
an
ever-­‐growing
corpus
of
informa3on
Low
accessibility
to
informa3on
and
data
Low
quality
of
informa3on
cura3on
and
lack
of
ability
to
integrate
disparate
informa3on
Ad
hoc
nature
to
managing
flow
of
informa3on
with
collaborators
and
partners
High
accessibility
to
informa3on
through
search
and
knowledge
dashboard
capabili3es
High
linkage
across
informa3on
and
datasets
A
single
simplified
external
partner
and
collaborator
pla?orm
for
informa3on
sharing

3. Our
approach
• Knowledge
is
a
shared
resource
that
we
need
to
collect,
evolve,
and
leverage
• Get
user
adop3on
one
itera3on
at
a
3me…and
do
Beta
releases
• Old
concepts
meet
new
technological
capabili3es…MDM,
Search,
Dashboards,
Graph
compu3ng,
Informa3on
Portal,
Informa3on
Hub
• 3-­‐clicks
to
any
informa3on
• Introduce
Content
Analy3cs

4. Data
Text
Connec3ng
IT
Users
world
Dashboard
Informa3on
challenge
Internet
+
Intranet
Internet
+
Intranet
Internet
+
Intranet
{Ownership,
Security,
Source}

5. ETL
&
ELT
Search
Pla?orm
Data/Informa3on
Hub
Social
Informa3on
Portal
Knowledge
Dashboard
Solu3on
&
Approach
{360
view}
Taxonomies
Dic3onaries
TrustedKeys
Lexicon
Ownership,
Security,
Lineage
Old
concepts
meets
new
technological
capabili3es
Apply
constant
transforma3on
Graph
Compu3ng
Content
Analy3cs
FingerPrin3ng
Crowd
Cura3on
Channeling
Profiling

6. ETL
&
ELT
Search
Pla?orm
Data/Informa3on
Hub
Social
Informa3on
Portal
Knowledge
Dashboard
Solu3on
&
Approach
Custom
built
framework
Taxonomies
Dic3onaries
TrustedKeys
Lexicon
Ownership,
Security,
Lineage
Technology
stack
Graph
Compu3ng
Custom
built
framework

7. Social
Informa3on
Portal
FingerPrin3ng
Crowd
Cura3on
Channeling
Profiling
Sugges3on
Algo
Creator/Consumer
Security
Lineage
Informa3on
at
your
finger3ps
Channel
Title
URL
Visual
User
Tags
Profile
Organiza3on
LinkedIn
+
AD
Interests
Index
Usage
fingerprin3ng
fingerprin3ng
Matching
(lexicon
+
Algo)
Feed
the
Search
Pla?orm
RSS
back
content
WorkSpace
Intent
1
1..*

8. Search
Pla?orm
Index
Document
Record
Linguis3cs
(Concepts
&
Similarity)
Language
of
Science
(Concepts
&
Vocabs)
Custom
NLP
TMA
Taxonomies
Dic3onaries
TrustedKeys
Lexicon
Faceted
Search
{en3ty,
{Document,
{Concepts}}}
Content
Analy3cs
API
query

9. First a 360 degree view
Right Molecule/Substance (SubstanceID)
Right Target (ProteinID,GeneID)
Right Tissue (CellType, CellLineID)
Right Delivery (Formulation vocabs, Device vocabs)
Right Patient (BioMarker vocabs)
Right Commercial (GeoCommercial vocabs)
Right Monitoring (Safety vocabs)
…
9Controlled
vocabularies
Doc
Data
En3ty
Graph
360
View
Vocabs
&
IDs
Content
&
Concepts
&
Provenance
Data
&
Concepts
&
Provenance
Visualize
Analy@cs
Aggregate
Knowledge
Dashboard

10. What makes up a Knowledge Dashboard?
Doc
Data
En3ty
Graph
360
View
Vocabs
&
IDs
Content
&
Concepts
&
Provenance
Data
&
Concepts
&
Provenance
Visualize
Analy@cs
Aggregate
Doc
Data
En3ty
Graph
360
View
Vocabs
&
IDs
Content
&
Concepts
&
Provenance
Data
&
Concepts
&
Provenance
Visualize
Analy@cs
Aggregate
Dashboard
En33es
network
Visualize
Analy@cs
Aggregate
Right Molecule/Substance
Right Target
Right Tissue
Right Delivery
Right Patient
Right Commercial
Right Monitoring
…all at once!
Custom
built
framework
…

11. Ac-on
points
4 things to take home
11
• Search
• Social
• Graph
• Lexicon
Search platform is your unstructured
content aggregator and more…
Leverage your community…have a
social information platform
A Information/Data hub is where you
connect it all…”data lake?”
You need a controlled vocabulary as
the metadata “glue”