Abstract
This paper introduces a new method to describe and analyse multidimensional time series based on wavelets. The methodology considers the time series as observations of a functional random variable. The paper generalizes previous research on stock market networks by including asset returns and volume trading as the main variables to study the financial market. The methodology is applied to examine the dynamics and structure of the Nasdaq-100 stock market during the pandemic period 2019/12-2021/12 considering both asset returns and volume trading to model the behaviour of different assets that are part of the index, applying an algorithm that offers better performance than others applied in the clustering literature. The study detects four clusters of firms corresponding with companies sharing common economic activities. The structure of the network reveals a nonlinear relationship between the variables, and the study shows that the main macroeconomic events during the period affect each cluster with different intensity. The change in the patterns of returns and risks and the redistribution of wealth in a highly changing environment are emerging phenomena, which must necessarily be carefully analyzed by public policies, in order to avoid the appearance of bubbles and systemic shocks.