The datasets currently used to calculate the average change in invoiceIDs. The average change from each month for each dataset is calculated separately and then these values are averaged to get a more stable line.