I have many sets of data, each data even within the sets have it's own date when the data first started and most recent to be uploaded into database table.
For example, we have 5 car companies (sets), each car companies produces many models, and each models have it's own data points. Each model is built on a different time, while some model of certain company already stopped production months ago, thus date range uploaded into database is not fixed.
When doing graph builder, the date is on an x-axis. I then fixed the date column's axis properties to make sure the increment for date is by 1 day. Some set started to have data for the past 5 months, while some only started to have within the last 2 months or few weeks. If I were to fix the column property for the axis, it will then take the overall oldest and most recent date in the database table. By doing this, I will have many blanks in graph builder (as shown in picture).
How do I make sure that what is displayed is only when data started and ends when it is last read? Not based on a specified range of minimum and maximum date? I cannot be doing subset on every group of data and concatenate again as it will not be efficient and script will be outdated and have to be redone when new group from the set is introduced into the database.
Accepted Solutions
Hello Azim-
Regarding your question about graphbuilder "displayed is only when data started and ends when it is last read?"
I am not sure about the structure of your data- but this example shows some of the options.
Graphbuilder will automatically determine the needed range of the X (time) access, based on the span of the data included in that graph pane.
Sample Data, Three car "model", with random "values" on "dates" in a Tall table.
For graph cosmetics- Dates for information about each "Model" do not need to be on fixed or aligned intervals (sample data made at 2 day, one week, and two week intervals)
When several Models are included in the same graph Pane- the Date axis will stretch to include all points. From your description this is not desired.
Overlay
Wrap, with levels per row forced to 1, is shown below. Using Model in "Group Y" would be similar.
But when "Model" is in the "Page" drop box- each graph is independent and will automatically set date span to show relevant data. If this is easily understandable by viewers, then it seems to answer your question. Note that time is on different scales, and it may be difficult for viewers to mentally adapt the shape of the profile to the individual timespan as they view several graphs together.
Another option to consider in situations where the Age (or time since model first released) is relevant- is to calculate an Age vs model introduction, and chart against that scale. This loses the Calander connection to model release date- but Time remains proportional and any trends of models relative to their release date remain viewable.
Hello Azim-
Regarding your question about graphbuilder "displayed is only when data started and ends when it is last read?"
I am not sure about the structure of your data- but this example shows some of the options.
Graphbuilder will automatically determine the needed range of the X (time) access, based on the span of the data included in that graph pane.
Sample Data, Three car "model", with random "values" on "dates" in a Tall table.
For graph cosmetics- Dates for information about each "Model" do not need to be on fixed or aligned intervals (sample data made at 2 day, one week, and two week intervals)
When several Models are included in the same graph Pane- the Date axis will stretch to include all points. From your description this is not desired.
Overlay
Wrap, with levels per row forced to 1, is shown below. Using Model in "Group Y" would be similar.
But when "Model" is in the "Page" drop box- each graph is independent and will automatically set date span to show relevant data. If this is easily understandable by viewers, then it seems to answer your question. Note that time is on different scales, and it may be difficult for viewers to mentally adapt the shape of the profile to the individual timespan as they view several graphs together.
Another option to consider in situations where the Age (or time since model first released) is relevant- is to calculate an Age vs model introduction, and chart against that scale. This loses the Calander connection to model release date- but Time remains proportional and any trends of models relative to their release date remain viewable.