The study looks at associations between music features and the weather (rain, sunshine hours, temperature) for over 20K songs in the top 100 UK charts over a span of 60+ year (up to 2019). The authors find strong weather associations for music features that reflect positive emotions and high intensity but not for those with low intensity, negative emotions. The authors also look at factors like popularity, with the leat popular songs in the chart showing no associations. Though correlational, the authors run several statistical tests to control for set variables, like seasonality, and they also address limitations of these in the discussion. They also run Generalised Addictive Models on justified PCA-ed data. Though I am not familiar with GAMs, the authors explain these sufficiently clearly to convey the rationale for these and for the reader to understand the outcome and implications of results. It will be interesting to see more studies like this one run on a wider sample of songs outside of the UK