3. EU Analytics: overview
3
Product lines
• Credit Bureau scoring, for the
financial and non financial industries
(i.e. Telcos and Insurances)
• Credit Bureau new solutions, such as
analytic based solutions (i.e. Financial
stress Index, Credit Limit, Profiling,
Credit Fair Value for risk assessment,
No hit scores, ……) and service
improvements
• Credit Bureau scoring preselling,
with analytics development and case
studies that show the benefits of using
credit bureau scores and data
• Eurisc Italy
• CB Czech Rep.
• CB Slovak Rep.
• CB India (High
Mark)
• CB Vietnam
(PCB)
• CB Jamaica
• CB Indonesia
• CB UAE
• CB Tajikistan
• CB Indonesia
• ……..
Clients
Product
management
Providers
• Product
Analytics Italy
• CRIF India
• ……..
4. What’s going on in Italy
4
Perform 2.0
• Developed in 2014
• Live from July 2015
• Economic impact analysis design
• One to one meetings based on
customized analysis: over 15
meetings already managed;
Perform Telecom
• Custom version realized for H3G
Perform Zone Consumer
• In progress
Perform Zone Business
• Kick Off within the year
Inserting EURISC in internal banks’
processes (POC on BPER, BPM)
5. What’s going on in Czeck & Slovack Rep.
5
CCB
• Promotion of Finantial Stress Index
live from October 2014
SCB
• Third generation of consumer credit bureau
score on banking and non banking bureau was
developed in the first mid of 2015 and it is live
form mid of June
6. What’s going on in Vietnam
6
PCB GLocal validation
• Market education started in 2013
• PCB has approved in July 2015
• Kick Off done in July 2015
• Go live for the end of 2015
PCB Credit Bureau score
• The first custom version of the PCB
credit bureau score is forecasted
for the end of 1016
7. What’s going on in India
7
High Mark
• MFI Credit Bureau Score delivered
in July 2015, will be in production
within 2015
• Zone score development in
progress
• No Hit Score development in
progress
• Commercial Credit Bureau score
development in progress
8. What’s going on in other countries
8
UAE
• Al Ethiad Credit Bureau score will
be developed starting from
September 2015
Azerbaijan
• GLocal validation completed
Jamaica
• GLocal validation in stand-by
9. General Considerations
Credit Bureau Solutions are becoming more and more crucial to defend
the value of data
In some case they could represent the marginality not reachable with the
simple data selling
Having the solutions often is not enough: it is necessary to show that they
bring real economic benefit
They will allow to integrate the new available information sources and
tools
So we work to improve
…. but we are ready to fly
9
20. We are here
The expected journey
New tool
New
methodology
Existing CB
data
Find new
data relations
or patterns
Build new
services for
customers
Existing CB
and new data
Find
additional
data relations
or patterns
Build even
better
services for
customers
Area of Great Responsibilities
20
21. The tool
3rd Party Technology platform
provider
The BDA product used is
MANTICA, running on Hadoop
framework
Usage of artificial neural network
to perform intelligent data mining
Currently used features:
– SOM alghoritm
– Graphs
What is an Artificial Neural
Network?
Computer programs implementing
sophisticated pattern detection and
machine learning algorithms on a
computer to build predictive models
from large historical databases
21
22. Methodology driven by the tool: SOM Algorithm
The SOM algorithm (Self-
Organizing Map) works on
data samples
It aggregates the samples
in clusters (codebooks)
whitin a map, upon a
“training” phase
It allows a quicker /
easier analysis of data
patterns or behaviors
22
23. Methodology driven by the tool: Graphs
We’re using two type of
graphs:
– Of records
– Of variables
The graphs represents the
items, grouped by the
strength (weight) of their
relations
It allows the identifications of
hidden groups of relations
It allows to show also sub
level of links (MRG)
23
24. What’s next
Maturity of the tool
WWCC1knowledgeoftool/methodology
Quick
wins
Patterns
identified
Services to
customers
New product still evolving (e.g.
enhancement in progress for
supervised analysis)
No shortcuts: clear definition of the
business problem and
understanding of data are key
It’s not a disruption vs. traditional
methods to analyze data, it’s rather
an evolution based on:
– knowledge
– curiosity
– objectives
– cooperation
– patience
Repetable
models
Alternative
analysis
24
An Overview of Data Mining Techniques (Excerpted from the book Building Data Mining Applications for CRM by Alex Berson, Stephen Smith, and Kurt Thearling)