Every data point and paper-metadata field, shown as a distribution. Categorical fields are counted by value; numeric fields are binned into a histogram.