Databases Reference
In-Depth Information
advisable to create a new dataset containing more balanced positive and negative
frequencies. The following steps accomplish this:
1. Use the parallel plot to create a subset of positive response observations.
2. Use the parallel plot to create a subset of negative response observations.
3. In the Control Center create a sampled subset of the negative response
dataset containing the same number of rows as the positive response
dataset.
4. Drag the sampled negative response set over the positive response set and
drop.
5. Select “Merge datasets”.
6. Right-click on the newly merged dataset; select “View/Edit names and
notes” to give it a more meaningful name.
Joining datasets
Frequently the data needed for an analysis is found in multiple datasets. For
example, a business may generate its own sales data that, for a complete
analysis, needs to be combined with population data from the census bureau. A
row by row combination of such data is known as an
equi-join
. It requires
common identifying attributes in each dataset used to match rows in one dataset
to rows in the other. To do this:
1.
open both datasets
2.
drag one dataset over the other and drop
3.
select “Join datasets”
4.
in the join dialog, identify the columns in each dataset that are to be used to
match rows.
Data Exploration
Dataset overview
In the Control Center, right-click on a dataset to view its summary statistics. The
summary includes: number of observations, attributes names (columns),
and corresponding data types along with summary statistics for each attribute.
For datasets with missing values, information on quantity and location of the
missing values is also summarized.
