29 Oct 2023

Big Data Analytics:

Transforming Financial Industries
Dian Ramadhani
School of Economics and Business
Telkom University
Agenda
The Big Data Origin (What and Where) Structured vs Unstructured Data
Pattern and Insight from Big Data Data-Driven Decision Making
Big Data Properties Big Data and its relation with AI, IoT, Smart Concept
Big Data and its relation with Data Artificial Intelligence & Learning Methodology
Analytics and Data Science Big Data Analytics Use Case in Financial Industries
 Komunikasi
Penyebaran informasi, identifikasi
hoax / fake news / hate speech,
identifikasi pemimpin opini.
Bisnis:
Segmentasi pasar, iklan
berdasarkan selera, perilaku
pembelian.

Kemasyarakatan:
Polarisasi kelompok, pengukuran
perilaku masyarakat.

Pendidikan & Huburan:

Sistem rekomendasi, konten yang
kontekstual dengan selera dan
Machine kebutuhan.
Learning
Computer Math and Kesehatan:
Science / Statistics Deteksi penyakit, pencegahan -
IT pengobatan - perawatan penyakit
Data
Science secara presisi.
Software Iraditional Transportasi:
Development Research
Self driving car, pemantauan
kendaraan real-time.
Domains/Business
Kowledge Perbankan:
Deteksi fraud, analisa resiko,
penilaian kredit, rekomendasi
investasi
 Big Data Algorithms & Big Data shape human/social behavior
Social Engineering (news & social media)
& Human to democratize, politicize, and openness (Arab Springs)

Behaviour
fear, manipulate, control (US presidential election, Brexit, etc)
Compete for Attention
Algorithmic
Behavior
& Matching
Recomender Engine
 Toward Massive Personalization
BIG DATA

More data sources More insights

Flexibility of Big Data technologies Big Data can provide a whole new set
allows the usage of both of information, in order to reach an
Profile
omni-comprehensive and multi-level
Internal and external data customer view Data
Sturctured and unstruktured data
Unstructured
Interaction Data Attitudinal Data
Customer contact notes Social Email Options
Customer media Chat transcription Preference
survey text Contracts
Call Center note Needs and Desire
Scanned document Web crawling Web analytics Market Research
Email Logo web In Person dialogues Social Media
Competitor External
Web Cust. Experience
scans

Internal External (credit and risk)

agency data Behavioural Data Descriptive Data
Marketing data Behaviaoral Product &
P&L Telephone Orders Attributes Data Orders
Customer directory Transaction Characteristics
Survey results Socio-demographic
Payment History Self Declared Info
Customer contact logs
Price benchmark Usage History Social Geo /
Name, Address details
comparisons
Transaction history Demographics info
Structured
 Data is Everywhere

LinkedIn Instagram
Understanding Pattern
Finding Topics
Provides Granularity
 Sentiment Mapping

Representation of newsrelated to culturally diversepopulation in Australianmedia - Usha M

Rodrigues, Michael Niemann, Yin Paradies
Personality Measurement
Understanding Polarity

Network of political blogs during the 2004 U.S. presidential election.

(Interian and Ribeiro, 2018)
Corruption Network Growth

Spain & Brazil Corruption Network - Martins et al. (2022)

Fake News Dissemination
Mobility Tracking
Text Network
Topic Modeling
Background Decade 2040s

of Big Data
Artificial Super Intelligence
(Bettter than Human)

Decade 2030s
Artificial General Intelligence
We generate huge amounts of data (from
(Human Level Intelligence)
UGC / mobile habit to machine generation
data/sensors/IoT) Decade 2020s
Our society leaves a massive digital Artificial Narrow Intelligence
footprint (so our behavior/attitudes) (Domain Base)
Finding unexpected patterns is so exciting
(also useful for predictive analytics) Decade 2010s
Big Data, Data Analytics,
In 2020s, we are entering AI era, where Machine Learning
we needs massive analytics efforts
Source: Super intelligence (Nick Bostrom)
 Big Data 5V’s Property

Volume

Variety Big Data Veracity Value

Velocity

Volume, Variety, and Velocity are the "essential" charateristics of Big Data
Veracity, and Value are the "quality" of Big Data
 Big Data, Data Analytics,
Data Science
Analytics
The discovery, Interpretation, and communication of meaning full patterns in data (wikipedia)
The process to uncover hidden pattern, unknown correlation, and other useful information that
can help organizations make more informed business decision

Big Data Data Science Social Computing

The Science to extract
Large, Fast, Complex Qualification of human/social
knowledge/pattern from data
the nV’s data behavior

MATH
Statistics
Discrete Maths

Source Insight
Information Theory
Combinatorics
Decision Theory

Review, opinion, historical Machine Learning

Market segmentation, information
data, conversation, network Econometriciants
Management
Data Vix Builders
Statistical
dissemination, fraud detection,
friendship, CCTV, Vlog, personalized adv, purchase
scientists Programmers
Actuaries
DATA
location tagging, etc. SCIENTIST behavior, brand awareness, etc.
Information
Business
Economics Systems
Finance Computer Science
Marketing SI Developers Software
Operations Data Analysis Engineering
Management System
Development
 Structured vs
Unstructured Data
Unstructured Data
Difficult to organise information using
traditional mechanisms

89%
Structured Data
High Degree of Organization,
such as a relational database

11%
 Moving from Data to Action

Data Facts and statistics collected together for reference or analysis.

Information Facts provided or learned about something or someone.

Information and skills acquired through experience or education; the

Knowledge
theoretical or practical understanding of a subject

Strategy A plan of action or policy designed to achieve a major or overall goal.

Data Driven
Decision Making
The goal of data science is improving decision-making, as this
generally is of direct interest to business
Statistically, the more data-driven a firm is, the more productive it is.
One standard deviation higher on the DDD scale is associated with a
4%–6% increase in productivity.

Benefit of Data Driven Decision Making

https://www.plutora.com/blog/data-driven-decision-making
 The CRISP-DM

CRISP-DM or Cross Industry Standard

Process for Data Mining is a framework that
has been commonly used to perform data
science processes.
The CRISP-DM life cycle consists of six
stages: business understanding, data
understanding, data preparation, modeling,
evaluation, and deployment.
 IoT Digital Twin
Uses

Big Data,
AI, IoT, Generates Creates
Smart
Concept
Employed

Big Data AI
AI & Learning Methodology

Artificial
Intelligence

Machine
Intelligence

Deep
Learning
 Predictive Analytics
How can we make
it happen?

Historical
What will Perscriptive
Analytics
1 Data
Predictive
Algorithms
Model
happen?

tion
Predictive iza
Why did it
Analytics ptim
happen? O h t
si g
re New
Fo 2
Value

What Diagnostic Model Predictions

Analytics Data
happened?

Descriptive ight
Analytics Ins

tion t Study the past to study the future (Confucious)

ma gh
for dsi
In Hin
Predictive Analytics is the art of building and
using models that make prediction based on
Difficulty pattern extracted from historical data.
 AI
Opportunities

Factors Deep Learning Machine Learning

Data Requirement Requires Large Data Can Train on Lesser Data

Accuracy Provide High Accuracy Gives Lesser Accuracy

Training Time Takes Longer to Train Takes Less Time to Train

Hardware Dependency Requires GPU to Train Properly Trains on CPU

Hyperparameter Tuning Can be tuned in Various Diffrent Ways Limited Tuning Capabilities
Big Data Value to the Entreprise
 Use Case in Financial Industries

Fraud Detection and Regulatory Compliance

Risk Management
Prevention and Reporting

Customer Feedback and

Customer Segmentation Credit Scoring
Experience Analysis

Predictive Analysis for

Market Sentiment Analysis Portfolio Optimization
Loan Defaults
 Credit Risk
Mr. Doty

Take Loan

Rp. 200.000.000

Risk Management Check

(Machine Learning)

Lower Credit Risk Higher Credit Risk

Get the Approval Doty's Loan

for Loan Application
Purchase Car may Rejected
Rp 250.000.000 Will Allocate Lower Money
 Robo Advisor

Porto A

Goal
Porto B
Risk

Porto C

Customer Sign up Customer input Machine Learning Robo Advisor Recommends on Customer
goals and risk Invested Portfolio for the Transfer Money and
customer Start Investing
 Fraud Detection

Begum et al: Credit Card Fraud Detection using Machine Learning

Srivastava et al: Credit card fraud detection at merchant side using neural networks
Crisis Prediction

Squartini et al: Early-warning signals of topological collapse in interbank networks