We clustered stock returns in the S&P 500 using k-means and DBSCAN, reducing dimensions and removing market effects as we went along. We used existing metrics and developed some new ones to evaluate our clusters - some things worked, and some things didn't, but we used the clustering algorithms to find stocks that behave remarkably similarly despite coming from entirely different industries. Our work can assist with portfolio diversification, pairs trading, and risk hedging in the financial sector.