Data Analytics for Estimating the Disposition Effect
Name
Tõnn Talpsepp
Abstract
The aim of the thesis is to design a data warehouse and develop data wrangling and feature engineering procedures that enable to store and transform stock market transaction data to conduct estimations of the disposition. The disposition effect is a probabilistic measure describing the behaviour of investors and is estimated using transaction level stock market data set. I develop the requirements for feature engineering, create a data model based on the star schema, provide detailed structure of the physical data model, and implement it using a relational database. I develop data transformation procedures, which include data generation procedures and financial calculation algorithms written in Java. The data transformation procedures enable to generate data for the data warehouse fact table and make various calculations for dimension tables. I run simulations to validate the suitability of the developed database structure and data transformation procedures. The simulations generate over three hundred million cases of stock market investment transactions in the fact table with hypothetical stop loss orders. The analysis of the simulation results show that the proposed data models and their implementation is appropriate for the task; and the developed data generation and calculation algorithms work as expected and enable to gain important new information about the disposition effect.
Graduation Thesis language
English
Graduation Thesis type
Master - Software Engineering
Supervisor(s)
Rajesh Sharma
Defence year
2022