Increase precision in pie chart Kibana - kibana

I have a series of categories, where some of them has a really low occurrence in the data set; but there still data from those categories. This is the data:
OS,Count
iOS,2119832
"__missing__",242326
"Windows Phone",122
Windows,28274
Ubuntu,18
Other,4170
"Mac OS X",6641
Linux,348
"Chrome OS",8
Android,12478135
You can see that there a lot of data for Android, iOS and __missing__ categories and very low numbers for others. This caused the pie chart to just show these three and other are just shown in the Inspect and Label hint:
How can I force the chart to show all the categories? May I change the chart type for this?

I used the Lens visualization and set the Stacked Chart as default. The little numbers are not visible at first sight, but counts are visible when you hover the mouse on each bar and I don't have to inspect the data to see all the categories and their counts.

Related

How to display items that make up a slice of a pie chart in Kibana?

I am attempting to create a pie chart that will display the health status of various systems. This is what the chart currently looks like:
Each data point, in addition to the possible errors messages, has a name in its data.
Is it possible to configure the chart so the system names are shown in the pie chart? In the tooltip for each slice would be preferable, but any solution will work.
You can Add (button at the buttom of the configuration) a second visualization, which will then add another layer around your first visualization.
This is an example for log.level and around it labels.url.full:

Why there is no "Top ..." in "Order Option in Kibana Visualization creation

I just started learning ELK, the data I am using has the following sample:
"YEAR","QUARTER","MONTH","DAY_OF_MONTH","DAY_OF_WEEK","FL_DATE","UNIQUE_CARRIER","AIRLINE_ID","CARRIER","TAIL_NUM","FL_NUM"
2016,1,1,6,3,2016-01-06,"AA",19805,"AA","N4YBAA","43"
2016,1,1,7,4,2016-01-07,"AA",19805,"AA","N434AA","43"
2016,1,1,8,5,2016-01-08,"AA",19805,"AA","N541AA","43"
2016,1,1,9,6,2016-01-09,"AA",19805,"AA","N489AA","43"
2016,1,1,10,7,2016-01-10,"AA",19805,"AA","N439AA","43"
When I am creating my first visualization, I want to have a pie chart to show the top stats of "CARRIER", in Kibana, below is what I see:
You can see from the left panel setting "Order", there is no "Top" in the dropdownlist, which causes the chart area shows one pie for the lowest stats of "CARRIER" which in this case is WN, and the rest is grouped into the second pie.
Doesn't make any sense here.
I watched some tutorial videos and there supposed to have more options in the dropdownlist like Top, which means for pie chart creation.
Am I missing anything here? How do I create the needed pie chart?
Thank you very much.
The "Size" parameter is actually the "Top X" you are looking for.
In your chart, it shows only 1 value, and it chooses the first value by order of ascending.
Change it to have more values.

How to make stacked histograms on kibana?

I have logs with different log levels and I want to design a date histogram like the one at the bottom of the picture :
At a particular date it should show say WARN logs with red and INFO logs above that with blue. Also I don't have log level indexed. But I do have INFO and WARN as saved searches.
As per my understanding you have a field wherein you can search for INFO & WARN and based on these searches, you want to build a stacked histogram as shown above.
The solution for the same is:
Click On Visualize Tab and Select Visualization Type as Vertical Bar Chart.
Select search source as From a New Search.
Then design your visualization by selecting buckets type as X-Axis and aggregation as Date Histogram with the required field such as #timestamp.
Then add sub-buckets as Split Bars and select sub-aggregation as Filters.
Mention the filter value as per search query. For example: log:INFO (In filter1) & log:WARN (In filter2).
You will get the resulting stacked Histogram. Also you can choose the color as required by clicking on the log:INFO label (towards the right of the histogram) and can choose your required color.

Pentaho Report Designer stacked percentage chart Y values

I'm new to Pentaho, and I have a small problem I cannot seem find an answer to. I'm running stock Pentaho 5.0.1 stable Report Designer. I have a bar chart that I have setup with a data source, everything there is showing/working correctly. I have changed the bar chart to a stacked percentage, and it stacks as expected. The problem is the values of the chart show from 0.0 to 1.0 incrementing by 0.1. Aesthetically I'd prefer this range from 0% - 100%, but don't see a way to change how it displays.
I see some solutions related to CCC/CGG, but I'm not running any plugins. This is just a out of the box Bar Chart. Playing around with y-axis settings does change the scaling/min/max values but none of that changes the visual display of the chart from the calculated range of 0-1 to is percentage counter part.
Thanks,
On the chart properties, go to Y-Axis and change the property y-tick-fmt-str (stands for Y ticks format string) to 0.0%.

Issue with legend coloring in an RDLC

I'm working on a report using a business object data set. The business object is really simple it has three fields (ColorName, ColorCode, Date). I'm making a bar chart using the year from the Date as the category, ColorName as the series field, and count of ColorName as the "data" (the top area in the designer) field. So the graph will show each year along the bottom with series bars of the colors. In the "Series Properties" I have the fill set to ColorCode and my color code values are HTML color codes including the pound sign. The bar graph is working perfectly. The bars are the correct colors. The legend however is inconsistent. My current graph has 5 colors and only 3 of them are showing the correct color in the legend swatch. Two of them are black but they shouldn't be. If I change the color code for a ColorName then the graph updates correctly but the swatch is still black. I have another graph that is a pie chart which is doing the same thing but the legend works correctly in that one. Anybody have an idea where I can look to resolve this legend color issue?
It looks like if any year had no series values then the legend would not work correctly. And it seems to only happen when you add a series to the report. The solution was to update my query to return counts instead of rows; including 0 counts for series that had no records. This way every series/category group had it's own record in the data set.
I have applied the same case for Rating Score 1 to 5. For each series I needed to color each bar in the Bar Chart according to the following factors (1:Poor, 2:Average, 3:Good, 4:Very Good, 5:Excellent).
Right click on any bar in the chart and select Series Properties.
Select Fill, Set background fill options - Fill Style to Solid.
Enter the below formula in the formula editor:
=IIF(Fields!Rating.Value=1, "Tomato", IIF(Fields!Rating.Value=2, "Orange", IIF(Fields!Rating.Value=3, "#cdf8cd", IIF(Fields!Rating.Value=4, "#7de67d", "Lime"))))
Hope this helps.

Resources