Movements | Analysis Sheets

Run macro SheetGenerator to add an analysis sheet to the An file.

As prompted by the macro, select the type of sheet (OA, CA, PA) and set a period for analysis.
When PA is selected, the PP sheet is also added. These two always work as a pair.
After the sheet added, run macro SheetProcessor. Charts will be created.

Original values in light-blue cells can be changed as needed.

Rerun macro SheetProcessor when changes are made. The charts are renewed accordingly.

The OA-0 sheet is created by macro An_Creator. It shows several timeline charts for the whole data period.

'OA' stands for "Overview Analysis". '0' conventionally stands for the whole data period in this system.
This sheet is intended to provide analysts with an overview of the data and animal movements.

Changes in the original values in light blue cells can be made, if needed.
However, the time period should not be changed for the purpose of this sheet.
To see details for a shorter period, add the OA sheet using macro SheetGenerator.

The OA sheet is used to create the timeline charts for a limited time period to see the details.

'OA' stands for "Overview Analysis".

Run macro SheetGenerater to add an OA sheet.

A time period and a sheet ID must be specified.

Then, run macro SheetProcessor.. Timeline charts are created.
Select variables to plot from the pull-down menu in the light blue cells.

Each chart can plot two variables. For each of them, evaluation time must be set.

For what these variables represent, refer to the 5th line in the Data sheet.

The evaluation time for HGT takes 0 irrespective of the selection.

The evaluation time for UTID takes the data time interval irrespective of the selection.

After changes are made, rerun macro SheetProcessor.. Timeline charts are renewed.
The following controls are also possible.

The maximum value for Y-axis of a chart is automatically set, but it can be set manually.

To change line colours in charts, change font colours of the corresponding variables.

To make lines in the chart thick, make the font of the corresponding variables bold.

To change the time period, change "From" and "To" date/time in the 1st row of the sheet.

To change the chart size, change the values in light-blue cells from "Chart Size" to the right.

The CA sheet generates autocorrelation and average-by-time charts.

'CA' stands for "Cycle Analysis".
The autocorrelation chart shows the presence or absence of periodicity in data.
The average-by-time chart shows the average level of the value by the time of day.

If periodicity is suspected, analyse the period using the CA sheet.
First, run macro SheetGenerater to add a CA sheet.

A time period and a sheet ID must be specified.

A certain length of time is required for valid analysis. More than 12 days is recommended.

Next, run macro SheetProcessor for analysis. Charts are created.
The values in the light-blue cells can be changed.

ET

SheetGenerater

The maximum of ET is 12 hours as analysis using a value greater than this is meaningless.
The ETs for HGT and UTID are always 0 and data time interval, respectively, for the following reasons.

Variable HGT is defined as the raw height data (instantaneous value). It cannot have a time length for evaluation.

Variable UTID is defined as the distance travelled during a unit time interval. Its ET is always the data time interval.

The ETs for variables NVF and NVI are 0 and data time interval, respectively, for similar reasons.

Variables NVF and NVI represent counted numbers and their averages are meaningless.

Those given in the 'average by time-of-day' chart are only dummies.

Autocorrelation: How are they plotted and what do they mean?
Auto-Correlation is the correlation computed from two identical time-series data.

The initial correlation between them is 1.0 as the values of every pair are identical.

This is the Y-value at 0 on X-axis.

Next, the correlation is calculated between the two, but mutually shifted by one data unit.

If the time interval is two hours, the calculation is based on pairs of data two-hour apart.

The result will be the Y-value at 2 on X-axis.

Next, the correlation is calculated between the two, but mutually shifted by two data units.

With two-hour intervals, the calculation is based on pairs of data four-hour apart.

The result will be the Y-value at 4 on X-axis.

This calculation is repeated up to 60-hour apart pairs to create the charts in this sheet.

Without periodicity, the correlation approaches 0 as the time gap increases.

Starting at 1.0, he correlation decreases and eventually stays around 0. If the data has periodicity, the correlation temporarily increases after that period.

When this happens, the reverse (or negative) correlation may appear at the half time of the period.
How to read the 'average by time-of-day' chart
The 'average by time-of-day' represents the data average separately calculated by the time of day.
If the data time interval is 1 hour, for example, 24 such averages are calculated for 24 hours.
When reading the chart, note the followings.

Noticeable unevenness in the 'average by time-of-day' does not necessarily mean periodicity.

Imagine that an animal, active day and night, occasionally uses a high area only during the day time.

This behaviour, if included in the data, could raise the average HGT or AHGT value during the daytime.

However, if the behaviour is irregular, the autocorrelation chartwon't show any periodicity.

Even if it is regular, the autocorrelation cannot detect the periodicity if it is longer than 3 days.

Thus, only behaviour repeated at intervals less than 60 hours can be detected in this system.

Noticeable unevenness might also be caused by large data error.

Especially when the analysis period is short, a large data error is easy to affect the result.

When a significant unevenness in an 'average by time-of-day' chart does not associate with periodicity in the autocorrelation chart, check the timeline charts in the OA or PA sheet to check what causes this discrepancy.

The PA sheet plots a time-space neighbourhood density (tsND) chart and some other charts.

'PA' stands for "Path Analysis". The data period is segmented based on the tsND patterns. Then, the segmentation is transferred to the PP sheet.

Run macro SheetGenerater to add a PA sheet. A PP sheet is added together.

A time period and a sheet ID must be specified.

Run macro SheetProcessor. The tsND and some Plus charts are created.
The Plus charts are listed in cells L2-U2.

They can be removed or added: use the pull-down menu for addition.

The values in the light blue cells can be changed.

SheetProcessor

When there are no charts at all, they are created according to the designation.

When there is no matching between a designation and a chart, the chart is renewed.

When all matchings hold, an optional menu appears. Select one from the menu.

Follow the instructions below for the creation of the tsND chart.

Initially, run the macro with cell F2 (evaluation time) and H2 (for evaluation radius) being blank.

The macro set the initial value for cell F2 according to the evaluation type in cell E2.

Time

Data Size

Time

Data Size

The macro works out the initial value for the evaluation radius in cell H2 according to the data.

Once the tsND chart is created, check the result and adjust these values as needed.

The evaluation time in cell F2 can be any integer multiple of the data interval, but it must be more than 3 times.

Any value for the evaluation radius in cell H2 is acceptable.

Several candidates in the pull-down menu, which the macro prepared, can be used.

If the tsND type in cell E2 is changed, the macro renews these entries.

If the second evaluation radius is entered in cell J2, the second tsND is plotted in the chart.

The dual-anchor chart uses the coordinates in the 2nd row of the PP sheet for anchors.

Initially, the anchors are positioned at the first and last data points, but they can be changed.

Enter the coordinates in the cells in the 2nd row: the change is immediately reflected in the map.

The anchor positions can also be designated by the date/time of data points.

SheetProcessor

In this case, the coordinates of the data points at these date/time(s) replace the entries in the PP sheet.

Once the tsND chart is created, the option for segmentation becomes available.

SheetProcessor

The macro identifies the two phases of movements (localised / travelling) based on the tsND.

As a result, the animal movement is expressed as the repetition of the localised & travelling phases.

The macro initially finds as many as these phases and then obliterates insignificant phases.

If a localised phase is obliterated, it becomes part of a larger travelling phase.

If a travelling phase is obliterated, it becomes part of a larger localised phase.

At the final stage of the removal, analysts' judge is requested as the distinction becomes subtle.

This final stage starts when the number of phases falls below the entry in cell X2.

If the number of segments are too few at this stage, increase the number in cell X2 and rerun the macro.

The tsND chart at this stage has a new line, representing the segmentation.

The descending and ascending line segments represent the localised and travelling phases, respectively.

The start of each segment is labelled by the segment ID number.

The macro finds possibly insignificant phases and shows the analyst their ID numbers.

In the chart, the line segments and dots neighbouring to either end of the phases are coloured red.

The analyst selects the phases to obliterate and return their ID numbers to the macro.

If satisfied with the result, the analyst ends the segmentation process.

After the segmentation, take the segmentation option, again, to continue pruning or start over from the beginning.

When the segmentation is finally completed, transfer the segmentation data to the PP sheet.

The segmented animal path is plotted based on the data from PA sheet. The final output is a kml file.

'PP' stands for "Path Plotting".

Sheet PP is added to the An file together with the PA sheet by macro SheetGenerator.

The sheet has two path charts.

The right one represents the whole data period of the An file; the left represents the sheet data period only.

Run macro SheetProcessor after the segmentation data is transferred from the PA sheet.

It will colour the segments of the path in the left chart and create a kml file.

Column C shows the time length of each path segment.

Its font colours correspond to those of the lines; the bold to thick lines; the italic to broken lines.

Change the colour and style of fonts and rerun the macro to change those of the lines.

The kml file shows segmented and coloured animal path on the 3D Google Earth map.

this page

The coordinates of points A and B on the left chart are used for the dual anchor chart in the PA sheet.

The dual anchor chart is used to indicate the animal position of interest along the timeline.

Its effectiveness largely depends on the position of two anchor points.

Point A is initially positioned at the first and point B at the last data point of the sheet data period.

The coordinates in the 2nd row of the sheet determine their positions.

Change the coordinates to make the dual anchor chart most effective.

The change is immediately reflected on the left chart. Check the positions.

When modifying the dual anchor chart in the PA sheet, ddate/time can also be used to designate anchor positions.

Then, the coordinates on the PP sheet are replaced by those of the data points at that designated date/time.

Cell area [L1:P3] is used for setting thresholds for separately colouring the properties of removal candidates.

The threshold values can be changed. Cell colours also can be changed.

L column: The threshold values for the angle that the two sides sharing the removal candidate point make.
M column: The threshold values for the length of the shorter of the two sides sharing the removal candidate point.
N column: The threshold values for the length ratio of the two sides sharing the removal candidate point.
O column: The threshold values for the multiples of unit time interval between the removal candidate and its predesessor.
P column: The threshold values for the multiples of unit time interval between the removal candidate and its successor.

GPS Data Analysis System: Movements

How to add and use the sheets for analysis

How to use the OA-0 sheet

How to use the OA sheet

How to use the CA sheet

How to use the PA sheet

How to use the PP sheet

How to use the cell area [L1:P3] in the LeapCheck sheet