v0.6.2

• Add n_clusters argument to bcn to specify the number of clusters to be found (for now,
  by using k-means clustering)
• Avoid division by zero in scaling the data