Here, we are exploring whether a countries continent or region of origin has an impact on who wins the world cup. We are also exploring historical trends for different continents and regions. We are defining contintinents and regions as specified in the countrycode_data data set in the countrycode library. For this analysis, We reduced the data to just matches where different continents and regions faced each other. And calculated the percent each continent won out of inter-continental matchups.
Investigating the trend for the Continents
Investigating the trend for the Continents since 1990
Linear Model
We try to fit linear model to data since 1990. We get following statistics for each continent.
Continent Estimate Std.Error Statistic P.Value
Africa 0.002508224 0.005469121 0.4586155 0.6657556
Americas -0.004976705 0.002575055 -1.9326598 0.1111109
Asia -0.004707792 0.008994748 -0.5233935 0.6230601
Europe 0.000637634 0.005261927 0.1211788 0.9082687
Oceania 0.03125 0.039692831 0.7872958 0.5754088
Hence, it there does not appear to be any predictive trend based on the continent that a country belongs to; none of the linear models are significant and the plots do not show any major effect either.
Investigating the trend for the Regions
Investigating the trend for the Regions since 1990
Linear Model
CONTINENT EFFECT
We try to fit linear model to data since 1990. We get following statistics for each region.
Overall, there does not appear to be some predictive trend based on the region that a country belongs to. A couple of the linear models were statistically significant, but the co-efficienct are very low and are likely not actionably large. The plots also do not show any actionable trends.