Data sets of any size have challenges: they need cleaning, and they can be incomplete or invalid. Open-box answers are often not uniform, because people spell, shorten and capitalize in different ways. Cleaning makes data uniform without changing their meaning, and a cleaning log records every choice. Filtering keeps only some rows, transforming changes every item, combining adds or compares data, and visualizing shows it in a chart. Cleaning choices can hide voices, because bias can enter at every level.