What is Time Series data?

To visualize Time Series data, consider the diagram below, which shows a group of polygon objects as they change over time. They are temporally seperate, but spatially overlapping objects.

If these three objects existed in the same GIS layer, (indeed, these are found in the Time Series Regime Polygons layer), they would stack up on top of one another and obscure one another's shape. The result would be a representation in which the true areas of the objects could not be accurately distinguished visually. Depending on the order in which the objects are stacked, which is determined only by the order they happen to appear in the GIS table, they might be shown partially distorted or completely obscured, as seen in the next image.

What we are accustomed to seeing as maps, are in fact geographic snapshots in time, or Time Slices. Time slices are simply layers of spatial objects for which the valid date is the same for all the objects. When we open up a road atlas, all of the features shown are considered to be current as of the publication date. To see historic layers of spatial information overlayed one upon the other is unusual in printed maps, and is always done with a very limited number of layers which can be distinguished from one another. For our Time Series data, we have no limit on the number of objects that may occupy or overlap the same space, therefore the visualization of the data, results in the sort of problems described above. Even if we remove the fill color of the polygons, we are still left with a completely ambigious series of overlapping boundary lines, known in GIS parlance, as "spaghetti."

To make visual sense of Time Series data it is expedient to extract a subset of the data which is valid for a particular time, in other words to filter out a single Time Slice from the "spaghetti" of objects. Because the smallest unit of time in the CHGIS datasets is one year--the temporal granularity--if we can select all the objects which were valid for a single year, we will effectively create a Time Slice for that year.

To get an idea of how this works, see the following diagram, in which the changing time series of polygons are represented with vertical bars, while their respective begin and end dates appear as labels in blue font . The polygons are representative of historical places that change at different times, so the records for each unique historical object have asynchronous begin and end dates.

For example, the pale jade colored polygon, has three historical instances shown on the diagram. The first instance is valid from time 1 to time 4, the second instance from time 5 to time 9, and the third instance from time 10 to time 29. By contrast, the tan colored polygon was first established at time 5, and took jurisdiction over part of the area formerly part of the jade colored polygon. In other words, the area of jurisdiction of the jade polygon was reduced, and the area that the jade polygon lost was placed under the jurisdiction of the newly created tan polygon. This is a typical example of the changes being tracked in the CHGIS datasets.

In the center column are shown Time Slices, which are a subset of the Time Series data. The objects in a Time Slice are valid for a single year. Therefore, by selecting only the objects which were valid for time 2, we obtain a Time Slice at time 2, shown at the bottom of the center column.

To select a single year from the Time Series data, use the following method (presenting sequentially for ArcView, ArcGIS, and MapInfo).

ArcView Method

1. add a Time Series layer to ArcView, select the layer to be filtered, and go to THEME | QUERY on the menu bar
   

2. double click on the field name [Beg_yr] and it will appear in the query text box
   

3. single click on the less-than-or-equal-to button and it will appear in the query text box
   

4. type in the year you wish to obtain and it will appear in the query text box, within the parenthesis
   

5. single click on the and button and it will appear in the query text box, along with following parentheses
   

6. double click on the field name [End_yr] and it will appear in the query text box, within the second parentheses
   

7. single click on the greater-than-or-equal-to button and it will appear in the query text box after End_yr
   

8. type in the same year you wish to obtain and it will appear in the query text box, after greater-than-or-equal-to
   

9. alternatively you may type the query directly into the textbox. Check to make sure the query matches this syntax exactly:
   ( [ Beg_yr ] less-than-or-equal-to filter-year ) and ( [ End_yr ] greater-than-or-equal-to filter-year )

10. when the query syntax is correct, click on the NEW SET button

11. return to the Map View window and you will see that the objects that were valid for the filter year will be selected (in yellow)
    

12. with the Map View window still active go to THEME | CONVERT TO SHAPEFILE on the main menu
    

13. in the Convert to Shape dialog, first browse to the folder where you want to create the new layer, then type in the new filename. click OK. you will be prompted to add the new Shapefile as a theme to the view. Click YES.
    

14. click on the checkbox of the original time series layer to turn it OFF. then click on the new layer to make it active and click on the checkbox to view the spatial data of the new layer in the Map View
    

15. click on the OPEN THEME TABLE icon to view the attributes
    

16. make sure that the values shown in the BEG_YR field are all lower than the filter year, and that the values in the END_YR field are all higher than the filter year. If this is true, you have successfully extracted all of the spatial objects current for the desired filter year into a new layer!