The new version of Azure Machine Learning has been released; with a desktop application as well, the Azure ML Workbench. One of the standout features for me is the ease in which data can now be quickly wrangled from a raw form into the correct form for a data scientist to start using.
Transformations happen in steps that can be reverted or entirely removed from the process. General time functions, string splits & regex and bucketing come pre-loaded into Workbench, but if you want to use something more specific you can write a function in Python that can be applied to your data. Data can be filtered out based on columns or split into multiple datasets using random sampling, perfect for splitting data into training and validation sets.
Workbench is also able to quickly plot data onto various different charts to see how transformations have changed the shape or distribution of data.
Workbench is a great tool for speeding up the data wrangling for those who maybe used to using Excel, R or Python for shaping data, or any other toolset. Workbench also has great model management abilities and is well worth checking out!