If the user selects only one circle, the presentation of the report is straightforward: one column per year, one row per kind of bird.
What should we display, then, if the user selects multiple circles? The answer depends on the intended use of the report.
If the report is intended as input to statistical analysis, scientific rigor requires that data be comparable. In particular, many researchers standardize CBC data as birds per party-hour, in an effort to mitigate the effects of varying effort across the years.
Some users (e.g., count and regional compilers) are interested more in broad questions, e.g., in how many years did this species occur at all?
In functional terms, these needs are served by presenting the resulting report in two different forms:
The split year presentation
presents the data for each circle in a separate column.
The column label is the combination of the year_no and year_key fields from
the effort table. For example, the column
label is “51-314” for year number "051", year key "0314 ". Leading
zeroes and trailing blanks are suppressed.
Before the report body we will display a copy of the regional page's primary index entries for all selected circles, so that researchers can quickly see where the data in the report came from: which circles were worked in which years under which published names.
The lumped year presentation combines the data from multiple circles into a single column labeled with the year number.
If that column happens to contain data from only a single
circle, the column label will be the same as in the split
presentation case: the concatenation of the year_no and year_key. Otherwise,
the column label will be the year number followed by the
number of lumped circles in parentheses. For example,
“88 (5)” would indicate that the column is
for year number 88 and combines data from five different
circles counted that year.
Here, then, are the cases for report presentation.
If the user selects only one circle, each column of the report will display the data for one year number in which the circle was worked.
If multiple circles are selected, the format depends on whether the user selected the split presentation.
For the split presentation, each column contains the data for a single circle in a single year number. Multiple columns for the same year number are grouped together, with the circles ordered by their year key.
For the lumped presentation, the format depends on whether the user selected the checkbox.
For a rigorous report, there will be a column for each year in which all the circles are worked. This may result in a report with zero columns.
If the user has not selected rigor, each column of the report will represent all the circles worked in a specific year number.
Let's look at a specific case of overlapping circles. Here are some lines from the regional index page for New Mexico:
Peloncillo Mountains, NM {31°44'N 108°55'W} [93-109]
Overlaps 59.2%: Peloncillo Mountains, NM {31°44'N 109°00'W} [74-92]
Peloncillo Mountains, NM {31°44'N 109°00'W} [74-92]
Overlaps 59.2%: Peloncillo Mountains, NM {31°44'N 108°55'W} [74-92]
Overlaps 2.3%: Portal, AZ {31°54'N 109°08'W} [73-98]
Overlaps 0.1%: Portal, AZ {31°55'N 109°08'W} [73-98]
Suppose the user selects all four of these circles: both Peloncillo Mountains circles and both Portal circles. Here are the different reports that the user may select.
If the user selects the split presentation, the column labels will be the year keys: 72-846, 73-906, 74-894, 74-922, 75-938, 75-970… Because the Peloncillos circle was not worked in year numbers 72 and 73, the first two columns show only the Portal data. For the remaining years, each year occupies two adjacent columns, one for the Peloncillos and one for Portal.
If the user selects lumped presentation but not strict comparability, the column labels will run 72, 73, 74, 75, …; each column will display the sum of whichever circles happened to be run that year.
With both lumped presentation and strict comparability, the column labels will run 74, 75, …, with columns only for those years when the number of selected circles was the same.
The party-hours row in the table header displays the sum of the party-hours of effort for all circles in that column.
When multiple efforts are combined in lumped presentation, how
do we combine census data from different efforts in the same
column? We can't just blindly sum the numbers, because some
may be count week numbers and some may be questionable
(flagged with the “q” field in
the census record). Also, what about
count-week records (flagged with the “plus” field)?
To make scattered records easier to pick out visually, we
display zero counts as a hyphen (-). In
general, up to four numbers may appear in one detail cell.
Example: “7, 2?, 8cw, 1cw?” would mean seven
regular records, two questionable, eight count-week, and one
record that was both questionable and count-week.
In any case, records flagged as either questionable or count-week will not be included in the statistics cells of the row.