These metrics serve specific tasks. All of them are off by default, except the TURF metrics, which are shown only in TURF analysis.
F.count (Feature Count)
F.count (ufc) counts the distinct codes of the variable in the VALUE drop zone among the respondents (or records) in the cell. Empty values are not counted, and weights do not apply. For example, with media channels in rows, countries in columns and the ad ID in VALUE, F.count shows how many different ads ran on each channel in each country.
Without a variable in VALUE, F.count is 1 for any non-empty cell. See Feature Count metric - F.count for a step-by-step example.
Top Label and First Label
These metrics put a category label into the cell instead of a number. DataTile looks at which category of the row variable each respondent in the cell has (or of the column variable, when metrics are shown on columns):
-
Top Label (
tlb) shows the label held by the largest weighted number of respondents in the cell: the most common answer. -
First Label (
flb) shows the label of the first respondent in the cell, in the order of the data file. It is meant for data where each cell holds one record, for example a database of products or stores.
On a row that is a single category, both metrics simply show that category's label. They become useful on nets and in dashboards, for example to show the leading brand in each region as text.
Prob Share
Prob Share (psh) is meant for probabilistic variables: numeric variables that hold a probability per respondent, such as the probability of reading an issue of a magazine, used in reach and frequency analysis. A variable becomes probabilistic when it is declared so in the codebook, see Variable types in DataTile.
When a probabilistic variable is in the table, DataTile counts each respondent with their probability instead of 1. Prob Share is that probability-weighted count divided by Vert Base:
Prob Share = Σ weight × probability ÷ Vert Base × 100
It is the average probability among the respondents in the column base. For example, four men with probabilities 0.8, 0.5, 0 and 0.6 give 1.9 ÷ 4 = 47.5%.
-
With the probabilistic variable in rows, Vert% shows the same value as Prob Share.
-
With the probabilistic variable in columns, Vert% divides by the probability-weighted base instead, while Prob Share keeps the respondent base.
-
Without a probabilistic variable in the table, Prob Share equals Vert%.
Prob Share is not a rotated-data metric; it works on any database.
TURF metrics
TURF analysis has its own metrics: Reach (_tt), the share of the base reached by an item on its own; Cumulative (_tc), the reach of the item together with all items ranked before it; and Unduplicated (_ta), the reach the item adds to the items before it. Count and Vert Base are also available in TURF mode. See TURF in DataTile.