We are working with transaction data that has millions of rows. Before building dashboards, we want to understand if there are recommended ways to prepare or model large datasets for better performance.