Hands-On Data Science for Marketing

Data cleanup

Before we can start building clustering models, there are five tasks we need to do to clean up our data and prepare it for modeling. The clean-up steps are as follows:

1. Dropping canceled orders: We are going to drop records with negative Quantity, using the following code:

        df <- df[which(df$Quantity > 0),]

2. Dropping records with no CustomerID: There are 133,361 records with no CustomerID and we are going to drop those records with the following code:

        df <- na.omit(df)

3. Excluding an incomplete month: As you might recall from previous chapters, the data in the month of December, 2011, is incomplete. You can exclude this data with the following code:

        df <- df[which(df$InvoiceDate < '2011-12-01'),]

4. Computing total sales from ...