ACCT3672 Accounting Analytics for Business Decision Making

• PLO1 : Business Knowledge
• PLO2 : Problem Solving

• PLO2 : Problem Solving

• PLO2 : Problem Solving

• PLO1 : Business Knowledge
• PLO2 : Problem Solving
• PLO4 : Teamwork

• PLO2 : Problem Solving
• PLO3 : Business Communication

• PLO3 : Business Communication
• PLO4 : Teamwork
• PLO5 : Responsible Business Practice

• PLO4 : Teamwork
• PLO6 : Global and Cultural Competence
• PLO7 : Leadership Development

• PLO1 : Business Knowledge
• PLO2 : Problem Solving

• Assessment 1: Online Quzzes
• Assessment 2: Seminar participation
• Assessment 3: Case Study
• Assessment 4: Assignment

• Assessment 3: Case Study
• Assessment 4: Assignment

• Assessment 1: Online Quzzes
• Assessment 3: Case Study
• Assessment 4: Assignment

• Assessment 1: Online Quzzes
• Assessment 3: Case Study
• Assessment 4: Assignment

• Assessment 3: Case Study
• Assessment 4: Assignment

• Assessment 3: Case Study

• Assessment 1: Online Quzzes
• Assessment 3: Case Study

• Assessment 4: Assignment
• Assessment 3: Case Study

Analytical mindset for accounting decision-making

In this lecture, we will introduce the analytical mindset essential for effective accounting decision-making, a concept we will consistently reinforce throughout the course. Our discussion will commence with a deep dive into the process of transforming business decisions into insightful analytical inquiries. Additionally, we will offer an extensive overview covering various types of accounting data and the diverse array of tools available in the realm of data analytics.

Mastering accounting data: Data storage, extraction and wrangling

Firstly, we explore the generation and storage of accounting data within business databases. Subsequently, we delve into techniques for extracting and manipulating this data in preparation for subsequent analytical tasks, leveraging tools such as SQL and Pandas.

Python basics for accounting analytics

We'll cover the installation and configuration process of Python and Jupyter Notebook. Additionally, a brief tutorial will be provided on fundamental aspects of the Python language, along with an overview of the primary analytics packages utilized throughout the course.

Descriptive analysis of accounting data

Throughout this week, we'll uncover valuable insights by delving into the hidden characteristics of data. We'll introduce descriptive analysis tools, including summary statistics, data aggregation, data visualization, and anomaly detection techniques. These skills not only help us develop a solid understanding of the data at hand, they also allow us to better form our hypothesis and develop our test plans.

Fundamentals of SQL and Pandas

We'll dive into the fundamentals of SQL for data management, alongside essential data wrangling techniques using Pandas. Special emphasis will be placed on merging multiple datasets and conducting data cleaning operations.

Performing data analyses and communicating findings

We'll present essential analytical methods for analyzing accounting data, focusing on the selection of appropriate methods and data for addressing specific inquiries. Additionally, we'll explore strategies for effectively communicating your findings and insights.

Descriptive statistics and data visualisation

We will cover the techniques for tabulating and visualizing summary statistics, utilizing pivot tables, and employing distributional statistics and visualization methods to identify anomalous accounting figures.

Fundamental analysis and quantitative investing 

Starting this week, we will explore prominent real-world applications of accounting analytics. Our this week focus will be on fundamental analysis, which involves utilizing accounting data to assess the financial performance and riskiness of publicly traded companies. This approach aids professional investors in identifying investment opportunities within stock markets. We will specifically delve into the construction of effective trading strategies grounded in straightforward accounting-based valuation metrics.

Basic machine learning algorithms in accounting analytics

We will delve into the process of conducting basic t-tests, logistic regressions, and linear regressions. Our emphasis will be on understanding how these algorithms can address various analytical inquiries and on effectively interpreting the outcomes derived from these analyses.

Recharge week--No class

Predicting financial frauds

The detection of financial fraud has been a longstanding challenge in financial regulation and auditing. In this segment, we introduce methodologies tailored for identifying fraudulent patterns within corporate financial reporting. We'll cover the selection of predictors, the specification and training of prediction models, and the evaluation of model performance. These techniques extend to many other classification tasks in accounting analytics.

Case study: Accrual anomaly

We will try to replicate a trading strategy based on the accrual anomaly, which is one the most famous and successful trading strategies developed from fundamental analysis.

Earnings forecasting

Forecasting public companies' earnings is a high-stake prediction game. We use regression analysis to detect key predictors of corporate earnings. We examine how to develop, train and validate new earnings forecasting models, as well as methods for evaluating the forecast performance. We also illustrate how a good earnings forecasting model may help investors develop successful trading strategies in the stock markets.

Case study: Predicting financial frauds using accounting numbers

We will explore whether a combination of accounting figures can serve as indicators for detecting financial fraud.

Analysing textual disclosures

We will introduce fundamental natural language processing (NLP) techniques and their application in analyzing corporate financial disclosures. Our focus will be on utilizing textual analysis to quantify abstract concepts such as managerial sentiment and companies' exposure to nonfinancial risks. These measures hold significant implications for financial forecasting and capital market outcomes.

Case study: Model-based versus professional analysts' earnings forecasts

We will develop an earnings forecast model and compare its properties with earnings forecasts produced by sell-side financial analysts.

Further Topics in Accounting Analytics

We will provide an overview of further topics in accounting analytics, such as neural network forecasting, ensemble learning, topic modelling, semantic analysis, and large language models. Although we do not expect to cover the implementation details of these state-of-the-art methods, we seek to develop a high-level intuitive understanding these methods.

Case study: Can managers hype stock prices

We will investigate whether managers can manipulate stock market perceptions by employing a more positive tone in disclosures, despite poor actual performance.