DSGenAI for Finance & Accounting

Curriculum

Module 1: Introduction to Data Science for Finance & Accounting (5 hours)

Topics Covered:

• Understanding data, types of data, data science tools, and its importance in finance and accounting.

• Key Concepts in Finance and Accounting: Introduction to financial statements, key metrics (ROI, NPV, etc.), and how they relate to data science.

• Data Science Tools: Python, Excel, Jupyter Notebooks.

• Data Visualization: Using Matplotlib, Seaborn, Plotly for visualizing financial data.

Case Study:

• Fraud Detection: Detect fraudulent transactions in credit cards or financial statements.

No-Code Tools:

• Tableau/PowerBI, Google Sheets, Google Colab

Hands-on Activity:

• Identify the pattern for finance & accounting and make decisions

 

Module 2: Data Collection and Cleaning (5 hours)

Topics Covered:

• Data Sources in Finance: Stock market data, economic indicators, company financials, and alternative data (e.g., social media, sentiment analysis).

• Data Preprocessing: Handling missing data, data normalization/standardization, dealing with outliers.

• ETL Process: Extracting, Transforming, and Loading data for analysis.

Case Study:

• Detect transaction anomaly: Find the pattern recognition of finance account holders.

No-Code Tools:

• Power BI, Google AutoML

Hands-on Activity:

• Treatment & detection of missing & outliers

 

Module 3: Exploratory Data Analysis (5 hours)

Topics Covered:

• Basic Statistics for Finance: Descriptive statistics, distributions, hypothesis testing, correlation analysis.

• Financial Ratios: Liquidity, profitability, and solvency ratios.

• Visualizing Financial Data: Line charts for time series data, histograms, heatmaps, scatter plots.

• Data Analysis with Pandas: Grouping, pivot tables, aggregation, and filtering.

Case Study:

• EDA analysis: Similarity & dis-similarity matrix

No-Code Tools:

• Google Sheets, Google Colab, PowerBI

Hands-on Activity:

• Descriptive statistics and decision making on the defaulter list using hypothesis

 

Module 4: Machine Learning for Finance (5 hours)

Topics Covered:

• Supervised Learning: Regression (Linear, Logistic), Decision Trees, Random Forest, Support Vector Machines.

• Unsupervised Learning: Clustering (K-means, Hierarchical), Dimensionality Reduction (PCA).

• Feature Engineering: Feature selection, creating financial indicators, and handling categorical data.

• Model Evaluation: Cross-validation, precision, recall, F1 score, ROC Curve.

Case Study:

• Credit scoring, fraud detection, risk management.

No-Code Tools:

• Google Sheets, Google Colab, PowerBI

Hands-on Activity:

• Train & Audit an ML Model for Fairness Using No-Code Tools

 

Module 5: Predictive Modeling for Financial Forecasting (5 hours)

Topics Covered:

• Credit Scoring Models: Logistic regression, SVM, and Random Forest for credit risk prediction.

• Stock Price Prediction: Using machine learning to predict stock prices based on historical data, sentiment analysis, or macroeconomic factors.

• Portfolio Optimization: Markowitz Efficient Frontier, Modern Portfolio Theory (MPT), CAPM, and multi-factor models.

Case Study:

• AI in Predicting Financial Performance & Promotions

No-Code Tools:

• Google Sheets, Google Colab, PowerBI

Hands-on Activity:

• Predictive models for finance, risk data management

 

Module 6: Financial Analytics and Risk Management (5 hours)

Topics Covered:

• Risk Management Models: Value at Risk (VaR), Conditional VaR, Monte Carlo simulations.

• Credit Risk Analysis: Default probability estimation, stress testing.

• Market Risk: Analyzing market exposure and potential loss using data-driven methods.

• Operational Risk: Predicting fraud, cyber risk analysis.

Case Study:

• Portfolio Management: Optimizing a portfolio using real-time data and predictive models.

No-Code Tools:

• Google Sheets, MS Excel

Hands-on Activity:

• Estimation and control limit of various risk

 

Module 7: Accounting Analytics (5 hours)

Topics Covered:

• Financial Statement Analysis: Predictive models for detecting irregularities in financial statements (fraud detection).

• Auditing Analytics: Using data science to improve auditing processes (e.g., anomaly detection in transactions).

• Cost and Budget Analysis: Forecasting and optimizing costs and budgets using data models.

Case Study:

• Economic Forecasting: Using machine learning to predict macroeconomic indicators (e.g., GDP, inflation rates).

No-Code Tools:

• Google Colab, PowerBI

Hands-on Activity:

• Cost, budget allocation of risk and finance

 

Module 8: Advanced Topics in Financial Data Science (5 hours)

Topics Covered:

• Deep Learning for Finance: Neural Networks for financial forecasting, sentiment analysis from financial news, and stock prediction.

• Classification & Clustering Processing: Analyzing financial news, reports, and social media for sentiment analysis.

• Blockchain and Cryptocurrency: Data science applications in blockchain analysis, cryptocurrency price forecasting.

• Algorithmic Trading: Creating automated trading systems using machine learning models and financial data.

Case Study:

• Classification & prediction: Using supervised / unsupervised model to predict the financial reports

No-Code Tools:

• Google Colab, MS Excel, Google Sheets

Hands-on Activity:

• Sentiment and voice data / transcripts analysis using AI models

Final Project: DSGenAI for Finance & Accounting

Project Theme:

Ensuring Transparency, Fairness & Accountability in AI-Driven Business Management

Problem Statement:

Businesses and organizations increasingly rely on AI and data-driven decision-making for hiring, credit scoring, marketing, customer interactions, and risk assessment. However, these AI systems often face ethical challenges such as bias, discrimination, privacy concerns, lack of explainability, and misinformation risks.

This project aims to identify, assess, and mitigate ethical risks in AI-based business management using data science, ML, LLMs, and Generative AI while maintaining fairness, transparency, and regulatory compliance.

Project Guidelines

1. Project Selection & Scope Definition

   • Choose one of the 10 business challenges related to ethical AI.

   • Define a research question and ethical objectives for the project.

   • Identify the key ethical risks and AI challenges to be addressed.

2. Data Collection & Preprocessing

   • Collect relevant datasets from open sources.

   • Perform data cleaning, handle missing values, and remove bias.

   • Conduct EDA to identify bias, fairness issues, and ethical risks.

3. AI & ML Implementation

   • Choose an appropriate model (Classification, Clustering, DS, Generative AI, or Fair AI).

   • Train and evaluate using no-code tools (Google AutoML, DataRobot, Azure ML).

   • Evaluate using accuracy, precision, recall, fairness metrics (AUC, F1, SHAP, etc.).

4. Generative AI for Responsible AI Practices

   • Analyze biases in LLMs, marketing content, and chatbots.

   • Use GenAI for document summarization, compliance audits.

   • Ensure ethical standards in all generated content.

5. Ethical Risk Mitigation & AI Governance

   • Use explainability frameworks (SHAP, LIME).

   • Propose mitigation strategies and regulatory recommendations.

6. Visualization & Final Report

   • Use Power BI, Tableau, or Google Sheets to present results.

   • Submit a structured report with insights, risk analysis, and audits.

Assessment Criteria (Total: 100 Marks)