From data analytic point of view, indexing should start with 1. When we analyze a data table, we always call the first row as the 1st row, or row #1, not row #0. It will be very strange to label rows as 0, 1, 2, 3, .... It may be fine for people with Computer Science background. But it would create so much confusion for almost everyone else...