Data Analysis of Financial Data and Trading Strategies

This is the code utilized in the seminar Data Decision Science of the University of Augsburg (Chair of Statistics, Spring 2020). Munich Re (Group) accompanied the project.

Background

It was investigated to what extent a cost-average effect can occur while following technical trading strategies. Looking at five different global markets, the investigations were based on common mean-reversion and constant-mix strategies. Due to data protection reasons, the respective results cannot be published on GitHub.

Python Scripts

exploratory_data_analysis.py

EDA of market data: first visualizations and data transformations (e.g. currency conversion).

main_functions.py

Numerous useful functions that were employed in this analysis. It can be seen as toolbox of this project. It is imported by signal_analysis_class.py.

signal_analysis_class.py

Forms a class module employing the above-mentioned toolbox main_functions.py. This step helped to automate the process of analyzing different strategies, markets and trading rhythms.

signal_analysis.py

Executes the analysis implemented in signal_analysis.py.

constant_mix_analysis.py

Analyzes if a cost-average can occur while following different constant mix strategies.

evaluation_constant_mix.py

Summarizes the results retrieved by constant_mix_analysis.py.

evaluation_signals.py

Summarizes the results retrieved by signal_analysis.py.

visualizations.py

Different visualizations of our results used in the final presentation.