In the three upper levels of the item tree (Root, Bh, SPD) the third tab in the right part of the window holds the filters for observations. Observations shown correspond to the item node selected. That means that clicking on an SPD will only show observations that belong to this SPD. To see all observations for the whole survey select the observations tab when the root of the tree is clicked.
The observations filter is a spreadsheet-like table where observations can be sorted and filtered according to several critiria. Per-column sorting and filtering is also available.
Observation filters
Notice that this table (as all other statistical tables) is read only. To actually change the observations shown you have to enter data in the items that contain them.
To interact with the table click on any of the headers and make a selection from the sort/filter menu.
Clicking on a header
The menu for each header allows you to:
Sorting acts the same for all columns. To sort multiple columns just sort the first column and then click another one and select from its menu a sort option too. To reset sorting on any column select the No sorting option. Sorted columns will have a small arrow next to their names. The biggest arrow is shown in the primary sort column (the first column that you sorted) and it will get significantly smaller in the other columns.
Sorting multiple columns
Directly below the sort options several filtering options exist. These are different for each column and depend on the type of data each column contains. There are 3 types of filters
If you want to see for example all observations for a specific month, click the "Month" header on the table and select for the menu the desired month. The table will update showing only the respective observations.
As with sorting you can filter multiple columns at the same time, or even have columns that are sorted,filtered or both at the same time. A column that is filtered will have its name between the > < symbols.
Filtering multiple columns
Notice that the filters are fully dynamic. Each time they apply to the data shown at the table and not all the observations at the dataset. If for example you select to see the high prices of observations by selecting the upper quartile on the price column and then you select the low quartile on brands you will get the least popular brands that belong to observations with high prices and not the intersection of observations with high prices and least popular brands.
Therefore it is important to keep in mind that the order that you filter several columns is significant and plays a big role in the presentation of the data. As you add filters in the table, you gradually decrease the data shown narrrowing it down according to your criteria.
To remove the filter from a column select the (All) option from the column menu.
See also: