3. | Fintech as Ecosystem - China vs USA vs Japan
Source: https://medium.com/wharton-fintech/fintech-in-china-an-introduction-6b11abd9cb64
Q: What comes to mind for
Singapore/SEA?
4. | Types of Analytics and Visualisation
Descriptive
•Shows what
happened
•Presents data and
used for reporting
•Well used
Source: Adapted from https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/an-executives-guide-to-ai
Predictive
•Anticipate what will
happen
•Using probabilities and
certain assumption about
the world
•Employed in data driven
organisations source of
insight
Prescriptive
•Provide recommendations
on what to do to achieve
desired outcomes
•Used by leading data-
driven and internet
companies as a source of
competitive advantage
5. | Data, is the new oil.
What makes big data HARD?
• Volume
• Variety
• Velocity
• Veracity
• Timeliness
• Classifications / Labelling
• Representativeness of data
• Preparation, Access
• Governance, Privacy and more…
Ahem. “Garbage in, garbage out”
Or some may say.
Data is the base material and evidence. We use tools like excels, R/Python, Tableau to help
us explore and analyse data - to ask questions to it directly.
6. | But Science & Understanding is the key
1. Frame problem / question
2. Gather Data
3. Prepare data
4. Create model (Statistics
analysis, Machine Learning
or Deep Learning)
5. Validate Model / Fine-tune
6. Deploy in real world
/ Report —> Business
outcomes
The (Iterative) data science process:
**If answer is not satisfactory, iterate steps 1 - 6
Data science:
• Helps us understand and analyse actual phenomena
• To predict, recommend and make decisions
• Uses the math and statistical
• Rigorous and iterative
7. | Who’s who in the data analytics
Business analyst
• Knows the business and derives business value
• Applies insight to improve revenue / operations
• But usually don’t know the math/tools/process
The “Boss” / “Mini HODs” / CIO
• Sponsors the project
• Asks the HARD / silly questions
• Needs to be educated (usually)
Data Engineer / Data scientist
• Coder person, sometimes eccentric
• Should be quite pedantic about the how the
algorithm and math works the real world is
always more complex and “takes more time”
• See the split of roles below:
8. | Linear Regression
Source: Adapted from https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/an-executives-guide-to-ai
Intuition:
• Strong straight line relationship between
the 2 variables.
Business case:
• Understand product-sales drivers such
as competition prices, distribution,
advertisement, etc
• Optimise price points
9. | Logistic Regression
Source: Adapted from https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/an-executives-guide-to-ai
Intuition:
• Extension of linear regression with binary
outputs (“Yes”/“No”)
Business case:
• Classify customers based on how likely
they are to repay a loan
• Predict if a skin lesion is benign or
malignant based on its characteristics
(size, shape, color, etc)
10. | Support Vector Machine (SVM)
Source: Adapted from https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/an-executives-guide-to-ai
Intuition:
• Non-linear high dimension optimisation
to split 2 classes.
Business case:
• Predict how likely someone is to click on
an online ad
• Predict how many patients a hospital
will need to serve in a time period
11. | Decision Tree
Source: Adapted from https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/an-executives-guide-to-ai
Intuition:
• Splits data features (eg. colour, height,
isEmployed) into decision nodes which
branches out
• Data can be numerical or categorical
• Easy to understand and interpret
Business case:
• Provide a decision framework for hiring
new employees
• Understand product attributes that
make a product most likely to be
purchased
12. | Random Forest
Source: Adapted from https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/an-executives-guide-to-ai
Intuition:
• Imagine chaining together the power of
many different decision tree to improve
accuracy
Business case:
• Predict call volume in call centers for
staffing decisions
• Predict power usage in an electrical
distribution grid
13. | Real World Finance Examples
What are some examples you can think of?
14. | k-means clustering
Source: Adapted from https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/an-executives-guide-to-ai and https://medium.com/datadriveninvestor/k-means-and-its-working-5c29a3d38523
Intuition:
• Putting (k) groups of data into the same
“set” by virtue of their similar
characteristics
• Not pre-determined by human.
• Not to be confused of KNN.
Business case:
• Segment customers into groups by
distinct characteristics (eg, age group)—
for instance, to better assign marketing
campaigns or prevent churn