Creating a Chart with multiple series and null values
7 August, 2013
Hello,
I am attempting to create a chart to ultimately show trend analysis over time. We have a couple of unique considerations:
1. The X axis is time, with no issues in consideration
2. The Y axis is the result set we are trending over the time as expected.
What I have is two series in consideration and want to show a line/combination chart with both in consideration.
Assume I have two series/measures with values in one series and one series/measure with values in another. In the opposing series those values are null.
So ideally I want one chart plotting both sets of series. Given that I have to create the report as a Cross Tab report to align the X and Y axis respectively, is there any way to accomplish this?
Thanks!
- Jay
I am attempting to create a chart to ultimately show trend analysis over time. We have a couple of unique considerations:
1. The X axis is time, with no issues in consideration
2. The Y axis is the result set we are trending over the time as expected.
What I have is two series in consideration and want to show a line/combination chart with both in consideration.
Assume I have two series/measures with values in one series and one series/measure with values in another. In the opposing series those values are null.
So ideally I want one chart plotting both sets of series. Given that I have to create the report as a Cross Tab report to align the X and Y axis respectively, is there any way to accomplish this?
Thanks!
- Jay
Hello Jay,
Combination charts can be built off column based reports and do not neccessarily need to be a cross-tab. To give us a clear idea of the problem, are you able to provide a screenshot of the chart and data, showing the troubles you are having?
Null values are typically not plotted on the graph or displayed as a 0, depending on the circumstances.
Regards,
Cameron
Combination charts can be built off column based reports and do not neccessarily need to be a cross-tab. To give us a clear idea of the problem, are you able to provide a screenshot of the chart and data, showing the troubles you are having?
Null values are typically not plotted on the graph or displayed as a 0, depending on the circumstances.
Regards,
Cameron
Hi Cameron,
I think that would help to best illustrate what I am seeing. In our case we have a dimension, benchmarkset, with multiple values we intend to display as the respective lines.
We also have two measures/series in consideration. In my case what happens is that one of the benchmarksets (dimension) for one of the measures is a flat line by intent. It's a calculated measure, but it's a flat line for the purposes of it's functionality within that measure when compared against the other benchmarksets/dimensions. (see example 1)
If I change the series/measure to reflect a separate measure with results, you see that two of the three benchmarkset lines disappear (expected) but the System Average line is now no longer straight. It moves over time because it is not a fixed value. (see example 2)
In our case the two measures in question populate results of calculations against the same data set but are separate by intention. We run a lot of reports relative to one another in that manner (Series/measure 1 vs. Series/measure 2)
But for this chart, ideally I want the System line in example 2 to be able to display with the Cohort lines in example 1. I'm not sure if this is possible because it's multiple dimenions (lines) and multiple measures (data points on the chart).
I'm happy to provide any additional context that'd help. Thanks!
I think that would help to best illustrate what I am seeing. In our case we have a dimension, benchmarkset, with multiple values we intend to display as the respective lines.
We also have two measures/series in consideration. In my case what happens is that one of the benchmarksets (dimension) for one of the measures is a flat line by intent. It's a calculated measure, but it's a flat line for the purposes of it's functionality within that measure when compared against the other benchmarksets/dimensions. (see example 1)
If I change the series/measure to reflect a separate measure with results, you see that two of the three benchmarkset lines disappear (expected) but the System Average line is now no longer straight. It moves over time because it is not a fixed value. (see example 2)
In our case the two measures in question populate results of calculations against the same data set but are separate by intention. We run a lot of reports relative to one another in that manner (Series/measure 1 vs. Series/measure 2)
But for this chart, ideally I want the System line in example 2 to be able to display with the Cohort lines in example 1. I'm not sure if this is possible because it's multiple dimenions (lines) and multiple measures (data points on the chart).
I'm happy to provide any additional context that'd help. Thanks!
I'm not sure that my attachments made it through in the previous post, so I'm attaching them again here.
Hi Jay,
Thanks for the info & screenshots, though I'm a little confused how this data is being displayed in the chart, as it doesn't seem to line up with your report data page.
In regards to the system average line, I'm interested to know how this is being returned in your report.
If you include a table with your chart, does it match up?
The report sql would also be useful.
From what I'm gathering it looks like you are trying to work out a dynamic reference line, which is fine, though I don't understand how this line changes just by selecting a different series in the chart, as changing the series should only change which line is displayed, and should not actually affect the data itself.
Are we missing something?
Thanks,
David
Thanks for the info & screenshots, though I'm a little confused how this data is being displayed in the chart, as it doesn't seem to line up with your report data page.
In regards to the system average line, I'm interested to know how this is being returned in your report.
If you include a table with your chart, does it match up?
The report sql would also be useful.
From what I'm gathering it looks like you are trying to work out a dynamic reference line, which is fine, though I don't understand how this line changes just by selecting a different series in the chart, as changing the series should only change which line is displayed, and should not actually affect the data itself.
Are we missing something?
Thanks,
David
Hi David,
How would I go about providing the Report SQL? Is that something I can extract from the app itself?
I'm trying to think of the best way to explain how the data is stored to help understand the context of the request against the mechanics.
In our case Example 1 is a benchmark value to measure average performance over time. In terms of calculation, two of the three dimensions (both Cohorts) are calculated as point of time calculations for each month, hence the moving lines in Example 1. The third dimension, System, is retained for benchmark purposes as a hard figure across time, hence the flat line.
In Example 2, the same data used to calculate the System average in Example 1 is displayed but is not a hard number retained over time. It is a point in time calculation of "Results" rather than displaying a consistent number across history.
For our purposes, I'd love to be able to display the two Cohort lines from Example 1, which are the moving lines, and the System line from Example 2 since it is moving as well.
I understand the nuance is there because of how we retain data on our side, but that's where I wanted to understand the mechanics on the YellowFin side to know whether I can ever obtain the type of chart requested.
Thanks!
How would I go about providing the Report SQL? Is that something I can extract from the app itself?
I'm trying to think of the best way to explain how the data is stored to help understand the context of the request against the mechanics.
In our case Example 1 is a benchmark value to measure average performance over time. In terms of calculation, two of the three dimensions (both Cohorts) are calculated as point of time calculations for each month, hence the moving lines in Example 1. The third dimension, System, is retained for benchmark purposes as a hard figure across time, hence the flat line.
In Example 2, the same data used to calculate the System average in Example 1 is displayed but is not a hard number retained over time. It is a point in time calculation of "Results" rather than displaying a consistent number across history.
For our purposes, I'd love to be able to display the two Cohort lines from Example 1, which are the moving lines, and the System line from Example 2 since it is moving as well.
I understand the nuance is there because of how we retain data on our side, but that's where I wanted to understand the mechanics on the YellowFin side to know whether I can ever obtain the type of chart requested.
Thanks!
Hi Jay,
The following post will show how to check the sql for the report :How to see the SQL used for your report
I guess all we're trying to do is figure out the raw data this report is using, and what calculations you have done in Yellowfin , then the results returned from Yellowfin. This does explain the report data page not tying in with the charts, as it sounds like you have a few different reports you are displaying.
If including a table in your report, it will also show what data the chart is using to draw the lines, which we might be able to insert into our own database.
What we've done in the past is get the raw data in a CSV, and then the report XML, we then created the database and imported the report, so this is always an option. If going down this path, you can email it through to support@yellowfin.bi and reference this post.
Regards,
David
The following post will show how to check the sql for the report :How to see the SQL used for your report
I guess all we're trying to do is figure out the raw data this report is using, and what calculations you have done in Yellowfin , then the results returned from Yellowfin. This does explain the report data page not tying in with the charts, as it sounds like you have a few different reports you are displaying.
If including a table in your report, it will also show what data the chart is using to draw the lines, which we might be able to insert into our own database.
What we've done in the past is get the raw data in a CSV, and then the report XML, we then created the database and imported the report, so this is always an option. If going down this path, you can email it through to support@yellowfin.bi and reference this post.
Regards,
David
Hi David,
You are understanding my point exactly. I do have two distinct reports I basically want to layer on top of one another.
In our case, the results we are displaying on the provided queries are self calculated. We are doing our calculations in our cube.
In my case, I'm trying to create a report that involves multiple dimensions and multiple measures. In one combination of dimensions and measures, the value(s) are flat lines because across time they are retained as a aggregate value.
Ex.
Benchmark method - 2, 3, 4, 5, 6 over 5 months is saved as 4 for all 5 months as it's the average across those months. (i.e. Jan - 4, Feb - 4, March - 4, etc.)
Results method - 2, 3, 4, 5, 6 over 5 months is saved as 2, 3, 4, 5, 6 in the respective months and the line moves as a result of those values. (Ex. Jan - 2, Feb - 3, March - 4, etc.)
In my screenshots provided, Example 1 (Measure/Series 1) uses the Results method for both Cohort lines and the Benchmark method for the System line.
In Example 2 (Measure/Series 2) the System value uses the Results method for retention. The Cohort lines do not populate values within this measure.
What I want to do is line up all of the Results method series onto one report. It involves two Dimensions and two series/measures because of where the respective data element lie.
Does this help more than providing the raw SQL since we are retaining the calculations on our end? Thanks!
You are understanding my point exactly. I do have two distinct reports I basically want to layer on top of one another.
In our case, the results we are displaying on the provided queries are self calculated. We are doing our calculations in our cube.
In my case, I'm trying to create a report that involves multiple dimensions and multiple measures. In one combination of dimensions and measures, the value(s) are flat lines because across time they are retained as a aggregate value.
Ex.
Benchmark method - 2, 3, 4, 5, 6 over 5 months is saved as 4 for all 5 months as it's the average across those months. (i.e. Jan - 4, Feb - 4, March - 4, etc.)
Results method - 2, 3, 4, 5, 6 over 5 months is saved as 2, 3, 4, 5, 6 in the respective months and the line moves as a result of those values. (Ex. Jan - 2, Feb - 3, March - 4, etc.)
In my screenshots provided, Example 1 (Measure/Series 1) uses the Results method for both Cohort lines and the Benchmark method for the System line.
In Example 2 (Measure/Series 2) the System value uses the Results method for retention. The Cohort lines do not populate values within this measure.
What I want to do is line up all of the Results method series onto one report. It involves two Dimensions and two series/measures because of where the respective data element lie.
Does this help more than providing the raw SQL since we are retaining the calculations on our end? Thanks!