The Profiler's descriptive statistics offer a variety of functions to analyze and explore data tables in a Teradata database:
- The Profiler's functions provide business insight.
- The Profiler uncovers data quality issues that can jeopardize the accuracy of any models that are based on the data.
- The Profiler isolates the data used in building analytic models. For example, outlying values may sometimes be excluded from a model; in other cases, these values might be required to solve a particular business problem.
- Some processes used in analytic modeling may require a certain type of distribution of data
In the Profiler's Descriptive Statistics analyses, NULL values are handled through the generated SQL aggregate functions. This SQL ignores the NULL value and adjusts the number of observations in its calculation. This provides a deletion of NULL values.
Teradata Profiler provides the following statistical functions to analyze data.
| Statistical Functions | Description |
|---|---|
| Values analysis | Counts the number of values for a given column or columns.
|
| Statistical analysis | Determines the following statistics for numeric columns.
|
| Frequency | Computes the frequency of column values and the frequency of values for columns in a single column list. Generates simple statistics for any other column within table. |
| Histogram | Determines the distribution of a numeric columns giving counts with optional overlay counts and statistics. |
| Text Field Analyzer | Analyzes raw data to find out actual data type. |
| Scatter Plot | Plots sampled values of two or three variables in 2-D |
| Overlap | Counts overlapping column values in combinations of tables. Finds key values in common between tables. |
| Data Explorer | Automates exploration of tables or views within an entire database. |