Research Motivation
This project started from a personal curiosity. I study data science with a strong interest in psychology and decision economics — the way mood, environment and macroeconomic pressure quietly shape the everyday choices people make. I wanted to move past intuition and look at whether those forces actually leave a measurable trace in real data. So I built a country-level monthly panel covering Germany, Spain, Turkey and the UK, and asked the question quantitatively: do sunshine, temperature and macro context shift consumer confidence and retail behavior in ways I can actually detect?
Research Questions
Does same-month sunshine correlate with consumer confidence?
Folk wisdom says sunny days lift mood. Part A tests this directly with Pearson and Spearman correlations between monthly sunshine hours and CCI for each of the four countries.
Does last month's weather predict this month's mood?
If weather effects need time to surface in survey responses, a lagged relationship should appear. Lag-1 correlation tests whether previous-month sunshine moves current-month CCI.
Do summer and winter consumer confidence systematically differ?
The most extreme weather contrast should reveal the strongest effect. A two-sample Welch t-test compares summer (Jun–Aug) and winter (Dec–Feb) CCI within each country.
Can non-linear ML models recover signals that linear methods miss?
Pearson catches only straight-line patterns. Random Forest, Gradient Boosting, Decision Tree and KNN test whether weather plus macro context predicts CCI and retail when the relationship is non-linear or interactive.
Is consumer perception or consumer behavior more responsive to weather?
CCI (a smoothed survey index) and retail volume (real spending) are two different windows into the same population. Comparing prediction performance shows which one weather and macro variables actually move.
Can countries be grouped by climate-driven consumer profiles?
K-Means and hierarchical clustering test whether the 480 country-months separate into meaningful weather-and-behavior archetypes — and whether those archetypes match real economic groupings.
Why Data Science?
I wanted this project to do more than just visualise the question. Answering it properly meant integrating four public data sources, making cleaning decisions transparent, running formal hypothesis tests, building predictive models and being honest about what the numbers cannot tell me. That full pipeline — not any single chart — is what convinced me the answer is more nuanced than the intuition I started with.