I would highly recommend sorting the initial load of the repository stats in descending order of occurances. With this, the user will see the files at top, that have the highest occurance statistik and are thus more likely to be refactored.