100% tevredenheidsgarantie Direct beschikbaar na je betaling Lees online óf als PDF Geen vaste maandelijkse kosten
logo-home
Samenvatting (FSA) Financial Services Analytics, (Kris Boudt, 2024) - 14/20!!! €11,48
In winkelwagen

Samenvatting

Samenvatting (FSA) Financial Services Analytics, (Kris Boudt, 2024) - 14/20!!!

 0 keer verkocht

Het document is een uitgebreide en gedetailleerde samenvatting van het vak Financial Services Analytics, gedoceerd door Kris Boudt in de opleiding Handelswetenschappen: Finance- en Riskmanagement aan de Universiteit Gent. Deze samenvatting omvat alle theorielessen, aangevuld met zorgvuldig uitg...

[Meer zien]

Voorbeeld 8 van de 157  pagina's

  • 25 januari 2025
  • 157
  • 2024/2025
  • Samenvatting
Alle documenten voor dit vak (9)
avatar-seller
MK2002
,PART 1: GENERAL TECHNIQUES FOR TRANSFORMING DATA INTO ACTIONABLE INSIGHTS

In part 1 we discuss the various steps in the data transformation process: (i) understand how the data can be
useful to solve the problem, (ii) implement the data transformation task in a code script that allows to
automate the analysis, (iii) evaluate the accuracy of the analytics and visualize results, (iv) scale and deploy the
tools as a business process in a financial services organization. We will compare traditional statistical tools with
robust ones, as well as techniques to deal with model uncertainty. The focus of the first part of the course is on
time series data.

LECTURE 1: INTRODUCTION TO THE COURSE


TOPIC 1: TIMELINESS OF A COURSE FSA
Steps in data analytics:

Business problem → Data problem → Data solution → Business solution.

Data analytics proceeds by first defining the core business question (the business problem), then formulating
its data‐specific challenges (data problem). Next, analysts develop a technical or algorithmic approach (data
solution) that resolves these challenges. Finally, the outcomes are brought back to the business context,
delivering tangible value (business solution). The diagram illustrates how these steps connect and highlights the
key questions: “What?” for the business problem and “How?” for the data solution, culminating in “So what
does it bring us?” at the business solution stage.

How data analytics can improve your businesses:

1. Identify the Business Problem.

• Clearly define the question or challenge that needs addressing (e.g., improving customer retention,
detecting fraudulent transactions, forecasting sales).
• This step answers the “What?” of the situation.

2. Translate into a Data Problem.

• Determine which data is relevant and how to frame the business question in technical/data terms.
• For instance, if the goal is to improve customer retention, you might focus on historical purchase data,
customer demographics, and support ticket logs.

3. Design and Implement a Data Solution.

• Acquire, clean, and explore the data to identify patterns.
• Build models or algorithms (e.g., machine learning) that can address the question.
• This step answers the “How?”—it is where analytics methods or AI techniques come into play to
generate insights from the data.

4. Deliver a Business Solution.

• Translate your analytical results back into actionable recommendations or automated systems that
provide real business value.
• Evaluate outcomes (e.g., improved retention, reduced fraud, better forecasts) to confirm that the
solution addresses the initial business problem.
• This data‐driven pipeline ensures that insights and recommendations meaningfully connect to
organizational goals.

1

,Why this trend towards digitalisation?

1. Technology has changed
2. Customers have changed
3. Regulation has changed
4. Competition has increased

The shift to digital solutions in finance and beyond is being fuelled by four key changes: rapid technological
advances, evolving customer demands, new regulatory frameworks, and intensified competition from diverse
market players. These combined forces push traditional institutions to adopt modern digital offerings.

Opportunity of the 21st century

1. Cloud computing

Allows users and businesses to tap into massive server infrastructures for hosting applications and storing data.
Instead of requiring on‐premise hardware, one can access powerful computing and expansive storage through
the Internet. This shift significantly reduces upfront investments and increases scalability. Major tech firms—
Google, Amazon, and Microsoft—offer powerful, on‐demand computing and storage through their cloud
platforms. Behind these services are vast data centres filled with servers, enabling organizations to run large‐
scale applications without owning the hardware themselves.

2. Abundance of data

Modern firms face an explosion in available data—enough to be termed “data obesity.” Hal Varian’s quote
highlights the shift from scarcity to abundance. Simultaneously, a Forbes article points out the shortage of data
science talent, emphasizing the challenge of turning massive troves of data into actionable insights.

3. Open source code sharing

The modern era makes code sharing and community collaboration far easier. Projects like Google Summer of
Code encourage new contributors, while platforms like Stack Overflow and GitHub enable widespread access
to open source software, best practices, and immediate peer support. This fosters rapid advancement in
software development and data analytics.

Lesson Learned:

“Give me six hours to chop down a tree and I will spend the first four sharpening the axe.”
—Abraham Lincoln

Abraham Lincoln’s quote teaches that spending ample time preparing (sharpening the axe) ensures greater
efficiency and better outcomes when executing a task (chopping the tree). In data analytics or digital
transformation, careful upfront planning often leads to superior and more sustainable solutions. This will be
the focus of this course and the exam. Make sure to always work in step processes in a well defined and
ordinary thought true way before starting with something.

The four major drivers of the trend toward digitalization in financial services:

1. Technology Has Changed.

• Rapid advancements in cloud computing, processing power, and data storage make it feasible to
collect, store, and analyse more and more data.




2

,2. Customers Have Changed.

• Consumers now expect personalized, seamless, and quickly accessible financial products. They are also
comfortable using digital platforms and devices, expecting 24/7 services. They embrace digitalisation
by showing high trust in tech firms which interact with a huge number of their digital devices and use
their data for corporate purposes.

3. Regulation Has Changed.

• Laws like PSD2 (in the EU) require banks to share payment account data with third‐party providers
(TPP) (with customer consent), opening the door to new players and technologies in finance which
lead to increased competition.

4. Competition Has Increased.

• Tech giants and fintech start‐ups are entering the market with specialized offerings (e.g., mobile
payments), putting pressure on traditional banks to keep up with new, innovative digital solutions.


TOPIC 2: INTRODUCTION IN R
For the basis concepts and steps you need to follow in R take a look at the slides!

Assigning values to an object:

In R everything is seen as an object. You assign a name and value to an object by using the <- “left arrow”

Data types in R

R has different data types

• Character: text
• Numeric: number
• Logical: Boolean: TRUE or FALSE

Data structures in R

R has different data structures:

• Vector: a one-dimensional array of numbers and characters
• Dataframe: multidimensional arrays
• Matrix: an array of real numbers

Selection of elements in an object

• For one-dimensional objects we only need to indicate the position within square bracket [ ]
• For two-dimensional object we need to indicate the row and the column with square bracket [,]




3

,LECTURE 2: FROM TIME SERIES DATA TO ACTIONABLE INSIGHTS

We will use data that everyone has (price data of stocks) and transform the data into actionable signals that are
useful for risk assessment and investment decisions. To do this we will need to transform the data which is key
in financial management. This will be done by 3 packages we use in R (xts, quantmod, and
performanceAnalytics).

• The price data we use will need to be transformed into yield insights.
• Some examples of transformations on or price data are here below:
• Subsetting: Considering only a smaller set of prices (f.ex. only the price data of the past year)
• Smoothening: We smoothen the price data by taking averages over local windows
• Differencing: difference the price data and compute return to estimate financial risk and reward of
investing in a stock.
• Time series aggregation: grouping time series data together. F.ex. daily to weekly, monthly or yearly
observations.
• …

These transformations are done repeatedly to the data (= workflow) and are best programmed as funactions.

• f.ex. Transformed_price <- f(price)


TOPIC 1: FROM DATA TO INSIGHTS USING FUNCTIONS
Functions in R

What does a function look like and how does it work?

General syntax of a function: Example: Power2 <- function(x){

out <- x^2

return(out)

}




1. Function Name (Purple, Top Left): The function name is an identifier that allows the function to be called
elsewhere in the program. It follows standard naming rules, meaning it should be unique and descriptive of the
function’s purpose.

2. Arguments (Blue, Top Right): The arguments (parameters) are variables that accept values when the function
is called. These values are used within the function to perform calculations or operations. The function can take
multiple arguments, and they are enclosed within parentheses ().

3. Function Body (Green, Middle): The function body contains the actual logic of the function. It consists of
statements that execute sequentially when the function is called. These statements can include calculations,
conditionals, loops, and other function calls.



4

,4. Return Value (Purple, Bottom Right): The return() statement ends the function execution and returns a value
to the caller. If a function does not explicitly include a return(), it returns the last evaluated expression by
default.

In R, functions are treated as data objects. One key advantage of using functions is that they contain reusable
code, which helps reduce workload and minimize errors.

A major benefit of R as a statistical computing environment is that it includes many predefined functions. These
built-in functions enable users to perform data manipulation, statistical analysis, and visualization without
having to write code from scratch.

Base functions can be classified as the following:

• Statistical functions: mean(), min(), max(), median(), sd(), IQR, summary(),…
• Graphical tools: plot(), hist(), qqplot(), boxplot(),…
• Object handling functions: read.table(), read.csv(), readRDS(), saveRDS(), write.table(),
write.csv(), write.zoo(), save(), load(), rm(), head(), tail(),…
• Probability functions: rnorm(), pnorm(), qnorm(), rt(), pt(), qt(), rcqhisq(), pchisq(),…

Functions in R are Shared Through Packages

R organizes functions into packages, which can be classified into two types:

1. Base R Packages

• Base R includes a set of packages that are automatically installed along with R. Examples: stats, untils…

• These packages provide essential functionalities and are automatically loaded whenever an R session
starts.

• Example of a base function:

o sd(x, na.rm = TRUE). Here the standard deviation is computed where the na.rm means do you
want to remove (NA, not available) values. When this is set to TRUE this means yes, when set
to FALSE, this means no.

o Function args() lets you discover the arguments of a function. Where args(sd) will give you
sd(x, na.rm = FALSE), with FALSE as default in the function.

2. External Packages

• Some functionalities are not included in Base R and require additional packages. Example: quantmod,
useful for quantitative modelling.

• These packages must be installed manually and loaded into the session each time a new R session is
opened.

Downloading price data using quantmod

In our Quantmod function, price data is collected from Yahoo Finance. Yahoo! Finance provides financial data,
such as stock prices and historical trends, which can be accessed through an API (Application Programming
Interface). An API acts as a bridge between a client application (e.g., a user’s laptop) and a database, allowing
data retrieval without manual downloads.



5

,The process works as follows:

1. The client app sends a request to the API (e.g., asking for Apple stock prices).
2. The API queries the Yahoo! Finance database and retrieves the relevant data.
3. The database stores financial data, which the API fetches.
4. The API returns the data to the client app, enabling users to analyse or visualize it.

With the quantmod package installed we can use the getSympbols function to get the OHLC (Open, High, Low
and close- price) data for a certain stock. Afterwards we can see our data class with function >class(Stock) and
plot our data now loaded in R with the plot(stock) function.

Example:




Note: 1. Adjusted close (last column) is the closing price after adjustments for all applicable splits and dividend
distributions. It’s the column that represents best the value evolution from the viewpoint of the investor. 2. An
xts data structure is numeric data ordered by time (format: YYYY-MM-DD).

An xts (eXtensible Time Series) object in R is a data structure specifically designed for time series analysis. It
extends the zoo package’s time series functionality and provides additional features for efficient handling and
manipulation of time-indexed data.


TOPIC 2: FROM DATA INSIGHTS TO USING TIME SERIES ANALYSIS
We now have OHLC price data (the one we imported from yahoo finance). This is already a transformation,
namely from transaction data to a summary value per day. Such transformations are called time series
aggregation

→ We could also summarize the data by taking averaging. This transformation is called time series
averaging
→ We could also work with the change in the price data. This transformation is called differencing.
→ We could also take a subset of the data. This transformation is called subsetting.

We do these transformations because they are more suitable for obtaining a certain financial insight.



6

, Transformation 1: Subsetting of time series data

Subsetting time series data involves extracting specific time periods from a dataset. In R, you can use index-
based selection with xts, zoo, or ts objects.

Example:




>dim function shows us the dimension of the matrix/data. Here we se 86 rows and 6 columns.

Now since a picture says more than a thousand words we can use a plot to visualise our data into one plot in a
meaningful way.

A meaningful way implies that our plot should show trends, the intraday information in the OHLC price per day,
connect price evolution with volume and differentiate between days with positive and negative returns.

The best way to do this is by use of a 2-panel plot. The top panel shows a candle stick chart, whilst the bottom
panel shows signed volume.

Signed Volume: Signed volume helps analyse buying and selling pressure in the market. It is calculated by
assigning positive volume when prices increase and negative volume when prices decrease.

• Positive signed volume: Indicates strong buying activity (more demand).
• Negative signed volume: Indicates strong selling activity (more supply).

This combination of price movement and trading volume helps traders understand market trends and make
informed decisions.

A candlestick chart is a financial tool that visualizes an asset's price movement over time. Each candlestick
represents a chosen timeframe (e.g., 1 day, 1 hour, or even 1 second) and displays the opening, closing, high,
and low prices within that period.

• Green Candlestick (Price Increase): The stock closed higher than it opened.
o The box (body) represents the range between the opening (bottom) and closing (top) prices.
• Red Candlestick (Price Decrease): The stock closed lower than it opened.
o Here, the closing price is at the bottom of the box, and the opening price is at the top.
• Wicks (Thin Lines Above/Below the Body): Indicate the highest and lowest prices reached during that
timeframe but where the stock did not close.

Candlestick charts provide a clear visual of price action, helping traders analyze trends and market sentiment.




7

Dit zijn jouw voordelen als je samenvattingen koopt bij Stuvia:

Bewezen kwaliteit door reviews

Bewezen kwaliteit door reviews

Studenten hebben al meer dan 850.000 samenvattingen beoordeeld. Zo weet jij zeker dat je de beste keuze maakt!

In een paar klikken geregeld

In een paar klikken geregeld

Geen gedoe — betaal gewoon eenmalig met iDeal, Bancontact of creditcard en je bent klaar. Geen abonnement nodig.

Focus op de essentie

Focus op de essentie

Studenten maken samenvattingen voor studenten. Dat betekent: actuele inhoud waar jij écht wat aan hebt. Geen overbodige details!

Veelgestelde vragen

Wat krijg ik als ik dit document koop?

Je krijgt een PDF, die direct beschikbaar is na je aankoop. Het gekochte document is altijd, overal en oneindig toegankelijk via je profiel.

Tevredenheidsgarantie: hoe werkt dat?

Onze tevredenheidsgarantie zorgt ervoor dat je altijd een studiedocument vindt dat goed bij je past. Je vult een formulier in en onze klantenservice regelt de rest.

Van wie koop ik deze samenvatting?

Stuvia is een marktplaats, je koop dit document dus niet van ons, maar van verkoper MK2002. Stuvia faciliteert de betaling aan de verkoper.

Zit ik meteen vast aan een abonnement?

Nee, je koopt alleen deze samenvatting voor €11,48. Je zit daarna nergens aan vast.

Is Stuvia te vertrouwen?

4,6 sterren op Google & Trustpilot (+1000 reviews)

Afgelopen 30 dagen zijn er 64796 samenvattingen verkocht

Opgericht in 2010, al 15 jaar dé plek om samenvattingen te kopen

Start met verkopen
€11,48
  • (0)
In winkelwagen
Toegevoegd