I am trying to determine what is the best statistical test for the dataset that was given to me as it is: small (n=1), repeated (2012-2014), and unbalanced (some years data was not achieved). The researcher was looking at fruit quality differences across four locations, repeating the measurements at the four locations from 2012-2014. I am assuming that I need to use a non-parametric test because the data is far from normal, variances non-homogenous, and sample size unbalanced, eliminating any robustness. Is the Friedman test appropriate, seems so given my online searching.

I appreciate any advice