I have a data set (let's assume it's students and scores) in Tableau.
I have used the Rank_parcentile() to rank them:
Rank_parcentile(sum([score]))
I then use the following code to group the students:
IF [percentile] > 0.8
THEN "Top 20%"
ELSEIF [percentile] <0.8 and [percentile]>0.5
THEN "Mid 30%"
ELSEIF [percentile]<-0.5 THEN "Bottom 50%"
This works very well and groups them as so. The problem is that sometimes I have too many students and they can't fit in a single graph reasonably.
Is there a way to divide the top 20% in one graph , the mid 30% in another and the bottom 50% in another?
Thanks in advance :-)
An easy approach is put percentile on the filter shelf. Show the filter, and then interactively switch between the three groups of marks -- all on the same chart
Alternatively, you can use a similar calculation to make a trellis chart.
Related
I'm trying to put together a story and I'm starting to create the first pages with graphics. The graphs are based on tables that are in S4/HANA and have the format as shown in the first image but with more fields, in each line there are weeks (repeated, although they do not appear like that in the photo) and that week will have a condition; "SI" or "no".
When creating a graph I can represent the amount (y-axis) against the condition and the week (x-axis) and it would look like this (2nd photo). In addition to this, I would like to add another bar per week that represents the TOTAL, the sum of condition = "SI" and condition = "NO". Is it possible to do this from SAP Analytics Cloud or do I have to rack my brain changing the lines of the source table adding a total there?
As a clarification, I do not need to add 2 different numerical fields or anything, I would simply like to add the SIand NO of the condition column, if on the y-axis I put volume that represents volume and if I put amount that represents amount, but I am looking for a third bar With the total, I don't know if I explained myself well.
Thank you very much in advance, if you help me solve this I would be very grateful, greetings
https://ibb.co/dcCKKzF
https://ibb.co/kDN3ZXp
Is it possible to add percentages to the chart legend, or is it the only way to put percentages to the chart, by adding the "Percentile Value Data"?
Thanks for your help in advance.
AFAIK the percentile value is usually something different, it only makes sense if the categories are based on intervals, e.g. given 7 billion living humans and measuring their age, the top 0.1 percentile are the 7_000_000_000 * 0.1/100 = 7_000_000 oldest living humans.
So, percentile values make sense for things you can measure, like age, income, length - in other words, numbers, or at least things which can be ordered.
You wouldn't use a pie chart then.
Percentile values don't make sense for things that cannot be measured like car labels.
Take a look at https://forums.opentext.com/forums/developer/discussion/48914/how-to-create-pie-chart-with-percentage-instead-of-values
M Williams writes there (I just repeat this here):
In the chart editor, you can go to the "format data" tab, then to the value series section, click on the "labels" button at the bottom, delete the "value data" option from the values window, then from the dropdown below the window, choose the percentage option and add it. This will show your pie chart labels in percentage.
I am trying to create a stacked bar chart between 2 dimensions (one being week, other is a string - more below) and 1 measure. I am new to Looker.
But in order to get the stacked bar chart I figure that I need to
Create a pivot
Transpose the data
Create stacked bar chart
I am stuck with step 2 and 3.
The above steps is what I have thought could help me arrive at the visualization I want but would appreciate if there is any other solution for this.
Scenario:
I have a digital marketing data with me, where I am trying to create a visualization on ‘Leads’ generated,
I have 2 dimensions, 1 measure
Lead Created Week - Week on which a lead was created
Reason2 - reason a lead was created or what reason attracted the user most
Leads - count of leads
I have plotted the below visualization where I have created a pivot of week with Reason2 and Lead count under it and have plotted it on a stacked bar chart (as below).
Where each bar is a Reason (X-axis) and each color is the Week dimension, leads (count of leads) is in the Y-axis
Even though this is help what would be helpful is having ‘Week’ on X-axis, ‘Leads’ (count of leads) on Y-axis and ‘Reasons2’ dimension as a legend.
There is also one catch i.e., there are more than 100 distinct values in ‘Reason2’ field but I need to pick on Top 10 Reasons (i.e., the reasons that generated the most leads), which in the above screenshot I have achieved by limiting the display rows to 10 as the measure was already sorted.
I tried to flip around the pivot by having reason in pivot and Week as dimension, but if now the pivot is sorted in alphabetical manner and not by value of measures (screenshot below), I need the top 10 reasons which most leads per week.
How do I achieve the desired visualization of having week on X-axis with leads on Yaxis but only the top 10 Reasons should be generated.
I am using Tableau to create a productivity progress bar of actual work done, versus the goal for that day. I have gone through numerous resources, but none seem to apply to only using 2 measures, basically putting the actual data measure turned into a percent within the goal measure. I have the chart working by using a calculated field of Missed Goal ([Goal]-[Actual]) but I am wanting into to look like a 100% stacked chart. Here is a picture of what I have, I just can't seem to get the last step expanding the bars horizontally to be a %100 chart:
Currently looks like:
Wanting it to look like:
Thank you for any help!
This is a good candidate for a few LOD Calculated Fields.
I will give an example based on the SuperStore dataset where [Profit] is a percentage of [Sales]. This can be extrapolated to your use case.
Because you have two dimensions on the view, both will be needed to be accounted for in the LOD calculations below:
//LOD Sales
{Fixed [Segment], Year([Ship Date]): SUM([Sales])} / {Fixed [Segment], Year([Ship Date]): SUM([Sales])}
We will always want [LOD Sales] to equal 1 for bar-chart display purposes.
//LOD Profit
{fixed [Segment], Year([Ship Date]): SUM([Profit])} / {Fixed [Segment], Year([Ship Date]): SUM([Sales])}
This will create the percentage of [Profit] to [Sales].
From there, you'll want to turn off Stack Marks so the bars overlap eachother:
Analysis Menu > Stack Marks > Off
And probably change the axis display to a percentage:
Right click [Measure Values] > Default Properties > Number Format > Percentage
For simplicity, your existing filters should be placed 'On Context' as to not interfere with the LODs. (Please read above linked article for reasoning and alternatives.)
Right click on Filtered Field > Add to Context
The end result should look like this:
I'd like to create a grouped bar chart with multiple measures by color. I have the following measures:
Clicks
Impressions
Revenue
Profit
I have a total of 2 web sites that I am collecting data from, so I'd like to have a different color for each measure and the columns to be side-by-side. Also, I'd like for them to be separated by site as well.
I see the stacked bar columns, but that's not the way I want to visualize this. Anybody have any ideas?
As far as I Understood, you have 4 measures, 1 dimension(Websites) with 2 websites names. Now, you need to show side by side bar, each column has two websites i.e., 4 columns with 2 sub columns each.
Drag Measure Names and Websites to columns shelf and Measure Values to row shelf.
Now, from Show Me shelf select side-by-side bars
If you get 2 columns with 4 sub columns, swap positions of Measure Names and Websites.
If you didn't get different colors for different measures, drag Measure Names to Color in Marks.
hope this is what you are looking for.
Just like Y.Prithvi said, you would use measure names and measure values.
(disregard the fields in the filters card, I filtered my data to match what you were trying to accomplish)
See example: