FIN306: Module Overview

FIN306 : Overview

This module develops your skills in financial data science, combining rigorous statistical foundations with practical Bloomberg Terminal expertise. The course website is the primary resource : all chapters, slides, labs, and quizzes are there, and the links in the schedule below point directly into it. You will learn to source, analyse, and model financial data using Python, while understanding the econometric principles that underpin modern quantitative finance.

A theme running through every week is one that matters more now than it ever has: intellectual judgement. Generative AI can produce analysis, summarise data, and write reports at speed. What it cannot reliably do is decide what matters in a specific context, weigh competing risks under uncertainty, or be held accountable for a conclusion. In financial services, those tasks will always fall to humans. The most employable data scientists in finance will not be those who can generate the most output; they will be those who can evaluate output critically, identify what is missing, and communicate a clear, defensible position. That capacity for considered judgement is the thread connecting every topic in this module, from data quality and bias to backtesting and machine learning. We teach tools, but the goal is the judgement to use them wisely.

Aims

  • Develop competence in financial data analysis using Python and Bloomberg Terminal
  • Build understanding of statistical inference and machine learning for finance
  • Apply data science methods to real-world financial problems with appropriate rigour

Core Themes

The module is organised around six interconnected themes:

  1. FinTech Foundations : Industry landscape and data infrastructure
  2. Statistical Inference : Bias-variance tradeoff, frequentist vs Bayesian perspectives
  3. Portfolio Analytics : Mean-variance optimisation, covariance estimation
  4. Cross-Sectional ML : From regression to ensemble methods (tree-based approaches)
  5. Backtesting & Validation : Walk-forward testing, overfitting, false discovery
  6. Sequence Learning : Time series ML, foundation models, computational law

Learning Outcomes

On successful completion, you will be able to:

  1. Source and prepare financial data using Bloomberg Terminal and Python APIs
  2. Apply appropriate machine learning methods for cross-sectional and time series problems
  3. Evaluate model performance with rigorous backtesting and validation
  4. Communicate findings through professional presentations and technical reports

Weekly Schedule

NoteModule Pacing

This is a technically demanding module. Week 1 establishes foundations that everything else builds on. Bloomberg Terminal labs begin in Week 2. Coursework 1 is a short recorded presentation due Week 6. Week 7 is a bridge session : CW2 scaffold preview and the conceptual shift from linear models to machine learning. Scaffold notebooks are released in Week 8.

Foundation Block (Weeks 1-4)

Week Topic Materials Assessment
1 Foundations of Financial Data Science Chapter · Slides · Lab 1 :
2 Data and Measurement Chapter · Slides Bloomberg labs begin
3 Time Series Foundations Slides · Lab 3 · Lab 3 Bloomberg CW1 brief available on Blackboard
4 Volatility Modelling Chapter · Slides · Lab 4 Homework · Lab 4 Bloomberg :

Core Methods Block (Weeks 5-7)

Week Topic Materials Assessment
5 Robo-Advisors & Portfolio Slides · Chapter CW1 topic confirmation deadline
6 Alternative Finance Slides · Chapter CW1 due: 6 March, 23:59 (recorded presentation + slides + risk register)
7 CW2 Scaffold Preview: From Linear Models to Machine Learning : CW2 orientation : choose your scaffold by end of week

Advanced Methods Block (Weeks 8-10)

Week Topic Materials Assessment
8 Cryptocurrency & Fraud Detection Slides · Slides CW2 brief + scaffold notebooks released
9 Factor Investing: From Fama-French to Gradient Boosting Slides · Lab 9 CW2 development (Scaffolds 1 & 2)
10 Backtesting & Validation Slides · Lab 10 CW2 development

Synthesis Block (Weeks 11-13)

Week Topic Materials Assessment
11 CW2 drop-in / scaffold support : CW2 development
12 Module Synthesis Slides · Lab 12 CW2 finalisation
13 : : CW2 due

Lab Structure

Week 1: Lab 1 (Data Science Primer) uses Colab notebooks only : foundation concepts before Bloomberg work begins.

From Week 2 onwards, labs follow a Homework → In-Class structure:

Component Tool Purpose
Homework Colab notebooks Learn concepts with accessible tools (complete before class)
In-Class Bloomberg Terminal Apply with professional data + instructor support

This structure ensures you arrive prepared for hands-on Bloomberg work, while making efficient use of our limited terminal access (~20 terminals). All homework labs use Bloomberg data (primary) or simulated data alternatives, avoiding unreliable free APIs.

Assessments

CW1: Responsible Data Science in FinTech (30%)

Due: 6 March 2026, 23:59 (Blackboard) · Full brief: Blackboard · Submission portal: CW1 submission and logistics

Recorded presentation (5 minutes) plus slide deck (6–8 slides) and data risk register (1–2 pages, 8–12 risks). Submit one ZIP file containing all three via the FIN306 Blackboard assignment by 23:59 on 6 March. Choose one of four FinTech applications : robo-advisor backtests, AI credit scoring, algorithmic trading, or fraud detection : and critically evaluate its data quality, selection biases, and responsible practice implications. Confirm your topic with the module coordinator by Week 5. Assessment focuses on analytical depth and use of Week 2 concepts, not business description.

CW2: Applied Data Science with Critical Reflection (70%)

Brief released: Week 8 · Scaffold notebooks released: Week 8 · Due: 8 May (see Blackboard) · Full brief: Blackboard

Choose one of four scaffold notebooks (tree-based methods, factor replication, walk-forward backtesting, or sequence learning) and write a 2,500-word reflective report addressing: method rationale, data quality decisions, results interpretation, limitations, and responsible use in FinTech contexts. The notebook is provided 60–70% complete; you fill in the TODO sections and write the report. Assessment focuses on critical reflection, not coding from scratch. CW1 data quality skills feed directly into the report’s data quality and limitations sections.

Resources

Bloomberg Terminal

Python Environment

  • Recommended environment name: fds (see the site’s setup guide)
  • Notebooks are hosted in a shared repository: GitHub
  • Can run locally or via Google Colab

Module Handbook

Full module details, assessment criteria, and academic policies are in the Module Handbook on Blackboard. Assessment briefs for CW1 and CW2 are also available there.

Contact: Professor Barry Quinn · b.quinn1@ulster.ac.uk