I'm looking for a data set (ideally a csv file) that contains more than 800 obs and more than 20 variables statistically significant to compare nearest neighboors, linear/logistic models, penalised linear/logistic models, trees, random forest models results such as quadratic error, misclassification error, ROC curves, AUC…
I took a look on kaggle's dataset best rated but I always have an issue (not enough variables, not enough variables statistically significant, not enough observations, name of the variables not specified so visualising them or analysing them is meaningless).
Any help would be highly appreciated