Big Data in FinTech
Mahmoud Jalajel | @mjalajel
Oasis500 FinTech Workshop
Agenda
● Intro
● Big Data: A bird’s eye
● Big Data applications in FinTech
● Case Studies
● Q&A
But first…
Mahmoud Jalajel
Big Data Engineer
& Data Scientist
Working with:
• Blue Kangaroo (e-commerce)
• Ligadata (big data for banks & telcos)
• Mawdoo3.com (Arabic Content)
• Jordan Open Source Association (Tech NGO)
• Telecommunication Regulatory Commission
Working on:
• Analytics Platforms
• Search Engines
• Recommender Systems
• Fraud Detection
• Current: Natural Language Understanding
Tech companies are leading FinTech
● Facebook Messenger will transfer money soon
● Amazon is experimenting with student loans
● WeChat has been transferring money for a while now
● E-Wallets by Jordanian telcos
● New ways to lend, invest, and donate online — championed by FinTech Startups.
● Creative methods to reach niche audiences
● Techs are winning with entrepreneurship, flexibility, resourcefulness, and speed!
Tech Patents by FinTechs
Let’s Talk Big Data
Big Data
Laying the pipes!
Ingestion
ETL Process
Big Data
Software
Data Analytics
Discover & Decide
- Drives instant value.
- Visualization and simple
analytics.
- Discovers basic patterns and
high-priority items to investigate
deeper
Power of
Visualizations
Data Science
Data Science is an interdisciplinary
field about processes and systems to
extract knowledge or insights from
data in various forms, either
structured or unstructured, which is a
continuation of some of the data
analysis fields such as statistics, data
mining, and predictive analytics.
— Wikipedia, of course!
Driving Value!
Data Science Functions
● Exploration, Analytics and Statistics
● Visualization and Relationships
● Prediction, AI and Machine Learning
● Understanding (and speaking) human languages
Exploration & Analytics
Visualization & Relationship Discovery
Predictions
Classification:
● Is this user a male or a female?
● Will this user repay this loan or not?
Regression:
● Probability of user clicking ad.
● Assign credit score for user.
Clustering:
● Google News: Topic Extraction.
● FinTech: What is in common across all of
my converging users?
Anomaly Detection:
● Gmail: Spam Detection.
● FinTech: Which of these transactions seem
fraudulent?
Prediction = Data Fitting
Revisiting The Big 3
Honorable Mentions
● The Internet of Things (IoT)
● Mobile platforms
● Security
● Cloud Computing
● Cryptocurrency
Big Data Applications in FinTech
Data Workflow
● Find the right dataset
● Collect and enrich data
● Data ingestion and organization
● Data visualization and exploration
● Data aggregation and analytics
● Discovering relationships
● Building predictive models
● Scaling models
● Personalization
● Contextual Personalization
Data Workflow
Find the right dataset Advisory services, opinion mining, build internal dataset
Enrich data with relationships Use open data sets
Data ingestion and organization Middle-tier technologies
Data visualization and exploration Middle-tier technologies
Data aggregation and analytics Analysing aggregate data
Discovering relationships Cybersecurity, Recommender Systems
Building predictive models Fraud, Process Optimization, Information security, cybersecurity
Scalability / Real-time problems Big data technologies
Personalization Personalized banking, tailored loan plans
Contextual Personalization Rented car insurance, travel health insurance
Big Data Impact on FinTech
Big data empowers:
● Solving traditional problem at a massive scale
● Integrating different data sources (user interactions, social profile, ..) in one
● Secure Transactions
● Real-time processing
● Access to historical data
● Discover patterns and relationships
● Automation of Workflows, Decisions and Alerts
Application Areas
● Solutions: Real-time processing
● Solutions: security
● Solutions: Data Analytics & Visualizations
● Solutions: Financial content aggregators
● Solutions: Credit score algorithms
● Automation: Micro loans, investment and insurance (by eliminating operators)
● APIS: Intermediate integration layers (phone to bank, bank to gov., bitcoin 2
creditcard ...)
● NLP: Social media opinion and sentiment mining
● NLP: Scout social space for user's profiles and reputation
The list goes on…
● ML/Classification: decide if the pledged user will realize the pledge
● ML/Regression: estimate most likely targets and stretch goals for a kickstarter
campaign
● ML: Advise on avoidable personal spendings
● ML: Risk assessment for investment opportunities
● ML: Fraud detection in banks and insurance companies
● ML/NLP: Cybersecurity for FinTechs
● RecSys: Personalized credit card and loan plans
● RecSys: General-purpose product recommenders for e-commerce websites
And on…
● Personalized Insurance Plans for Insurance Sector
● Minimize data/money Loss at major online retailers and banks
● Fraud Detection in banking and insurance industries
● Information Security at banks and government agencies
● Cyber Security for banks and government agencies
Tech Monetization
Case Studies
Liwwa
liwwa.com
- Crowdfunding/lending
- SME Financing
- The application/review/funding
process is fully automated.
- Using Machine Learning to
Decide on loans eligibility
(almost as good as an analyst)
- Automated risk assessment
Riskopy
riskopy.com
- Risk management
- Using graph databases to
discover hidden relations and
predict future events
- Using real-time processing to
monitor company’s activities and
send alerts
Financial Industry
Applies to e-commerce and other
domains as well.
Why do we need to analyse Social
Media data?
● More than 2b social media users
(30%
● Research shows that majority of
customers expect their requests
solved within 24 hours, if not
less than 5mins
● One key barrier found here is
analysing this data and extracting
insight at scale, and at the right
time
Major Banks
Use-case
How about you?
How do you envision using data in your startup?
Thank You!
Mahmoud Jalajel
@mjalajel

Big Data in FinTech

  • 1.
    Big Data inFinTech Mahmoud Jalajel | @mjalajel Oasis500 FinTech Workshop
  • 2.
    Agenda ● Intro ● BigData: A bird’s eye ● Big Data applications in FinTech ● Case Studies ● Q&A
  • 3.
  • 4.
    Mahmoud Jalajel Big DataEngineer & Data Scientist Working with: • Blue Kangaroo (e-commerce) • Ligadata (big data for banks & telcos) • Mawdoo3.com (Arabic Content) • Jordan Open Source Association (Tech NGO) • Telecommunication Regulatory Commission Working on: • Analytics Platforms • Search Engines • Recommender Systems • Fraud Detection • Current: Natural Language Understanding
  • 5.
    Tech companies areleading FinTech ● Facebook Messenger will transfer money soon ● Amazon is experimenting with student loans ● WeChat has been transferring money for a while now ● E-Wallets by Jordanian telcos ● New ways to lend, invest, and donate online — championed by FinTech Startups. ● Creative methods to reach niche audiences ● Techs are winning with entrepreneurship, flexibility, resourcefulness, and speed!
  • 6.
  • 7.
  • 9.
  • 10.
  • 11.
  • 12.
  • 13.
    Data Analytics Discover &Decide - Drives instant value. - Visualization and simple analytics. - Discovers basic patterns and high-priority items to investigate deeper
  • 14.
  • 15.
    Data Science Data Scienceis an interdisciplinary field about processes and systems to extract knowledge or insights from data in various forms, either structured or unstructured, which is a continuation of some of the data analysis fields such as statistics, data mining, and predictive analytics. — Wikipedia, of course! Driving Value!
  • 16.
    Data Science Functions ●Exploration, Analytics and Statistics ● Visualization and Relationships ● Prediction, AI and Machine Learning ● Understanding (and speaking) human languages
  • 17.
  • 18.
  • 19.
    Predictions Classification: ● Is thisuser a male or a female? ● Will this user repay this loan or not? Regression: ● Probability of user clicking ad. ● Assign credit score for user. Clustering: ● Google News: Topic Extraction. ● FinTech: What is in common across all of my converging users? Anomaly Detection: ● Gmail: Spam Detection. ● FinTech: Which of these transactions seem fraudulent?
  • 20.
  • 21.
  • 22.
    Honorable Mentions ● TheInternet of Things (IoT) ● Mobile platforms ● Security ● Cloud Computing ● Cryptocurrency
  • 23.
  • 24.
    Data Workflow ● Findthe right dataset ● Collect and enrich data ● Data ingestion and organization ● Data visualization and exploration ● Data aggregation and analytics ● Discovering relationships ● Building predictive models ● Scaling models ● Personalization ● Contextual Personalization
  • 25.
    Data Workflow Find theright dataset Advisory services, opinion mining, build internal dataset Enrich data with relationships Use open data sets Data ingestion and organization Middle-tier technologies Data visualization and exploration Middle-tier technologies Data aggregation and analytics Analysing aggregate data Discovering relationships Cybersecurity, Recommender Systems Building predictive models Fraud, Process Optimization, Information security, cybersecurity Scalability / Real-time problems Big data technologies Personalization Personalized banking, tailored loan plans Contextual Personalization Rented car insurance, travel health insurance
  • 26.
    Big Data Impacton FinTech Big data empowers: ● Solving traditional problem at a massive scale ● Integrating different data sources (user interactions, social profile, ..) in one ● Secure Transactions ● Real-time processing ● Access to historical data ● Discover patterns and relationships ● Automation of Workflows, Decisions and Alerts
  • 27.
    Application Areas ● Solutions:Real-time processing ● Solutions: security ● Solutions: Data Analytics & Visualizations ● Solutions: Financial content aggregators ● Solutions: Credit score algorithms ● Automation: Micro loans, investment and insurance (by eliminating operators) ● APIS: Intermediate integration layers (phone to bank, bank to gov., bitcoin 2 creditcard ...) ● NLP: Social media opinion and sentiment mining ● NLP: Scout social space for user's profiles and reputation
  • 28.
    The list goeson… ● ML/Classification: decide if the pledged user will realize the pledge ● ML/Regression: estimate most likely targets and stretch goals for a kickstarter campaign ● ML: Advise on avoidable personal spendings ● ML: Risk assessment for investment opportunities ● ML: Fraud detection in banks and insurance companies ● ML/NLP: Cybersecurity for FinTechs ● RecSys: Personalized credit card and loan plans ● RecSys: General-purpose product recommenders for e-commerce websites
  • 29.
    And on… ● PersonalizedInsurance Plans for Insurance Sector ● Minimize data/money Loss at major online retailers and banks ● Fraud Detection in banking and insurance industries ● Information Security at banks and government agencies ● Cyber Security for banks and government agencies
  • 30.
  • 31.
  • 32.
    Liwwa liwwa.com - Crowdfunding/lending - SMEFinancing - The application/review/funding process is fully automated. - Using Machine Learning to Decide on loans eligibility (almost as good as an analyst) - Automated risk assessment
  • 33.
    Riskopy riskopy.com - Risk management -Using graph databases to discover hidden relations and predict future events - Using real-time processing to monitor company’s activities and send alerts
  • 34.
    Financial Industry Applies toe-commerce and other domains as well. Why do we need to analyse Social Media data? ● More than 2b social media users (30% ● Research shows that majority of customers expect their requests solved within 24 hours, if not less than 5mins ● One key barrier found here is analysing this data and extracting insight at scale, and at the right time
  • 35.
  • 36.
    How about you? Howdo you envision using data in your startup?
  • 37.