You pair down the data to avoid redundant features and prune your data set until your clustering algorithms start making some splotchy plots with all your panda data categorizing into, well, you’re still not quite sure what those categories were, but they look like they’re clustering other than a handful of datapoints of oddballs just sitting in between whatever Panda sexual compatibility spectrums the clustering was fitting to. What do you do?
Try Another Algorithm—seems sketch this unsupervised stuff
Evaluate the results first, maybe it’s good, but should check
Trust the algorithm, it’s statistical distance, how are ya even gonna evaluate that thing?