As a reminder, we're still in the create the dataset phase of our ML model journey. In parallel with your feature engineering process, we want to explore and process our data before we feed it into our model. Now, let's examine some of the ways that we can do that. Recall from the first course in the specialization, where you're introduced with a few ways to explore your datasets. Now, I'll provide a link to the lesson video, so we'd have to go through the SQL basics again here. Now, what I do when I mentioned are two quick points. Number one, the dataset creation process itself is iterative. The ML model building process does not mean that we explore clean and transform the data table into BigQuery, and then we're done with all of our model inputs. You'll often find yourself iterating back to your dataset to creating clean new features, filter out bad values, and retry your model to see if your performance improves. Number two, this list is not exhaustive. If you're looking to become a data scientist in the future, you'll find other ways to explore distributions and common data values likely using IPython notebooks with Pandas DataFrames or using something from R in an R library. So, all I recommend is start with an approach that works for you and always remain skeptical and curious about your dataset. Now, let's do a quick demo on some advanced features in Cloud Dataprep to help us explore e-commerce dataset.