7 Replies Latest reply on Apr 26, 2012 6:16 PM by Dimitri.B

[solved] how to create chart with X axis: value ranges; Y axis: % of case (value >= a certain threshold value)

I have a data set with two columns and about 200 rows (csv file attached):

I'd like to create a chart as shown below:

May I ask how to achieve this in Tableau?

Thanks a lot!

J

• 1. Re: [help] how to create chart with X axis: value ranges; Y axis: % of case (value >= a certain threshold value)

See attached (doesn't look like your example, but that's what the data says).

The steps are:

1. Create 'brightness' bins (right-click -> Create Bins...)
2. Group 'satisfaction' scores into positive and negative (right-click -> Create Group...)
3. Build a  graph as laid out in the attached
4. Set 'quick table calculation' for SUM(Number of Records) to Percent of Total (click little down arrow on the right of the 'pill'), then set 'Compute using' to 'satisfaction (group) '
5. Hide negative section of the graph (right-click  -> 'Hide')
• 2. Re: [help] how to create chart with X axis: value ranges; Y axis: % of case (value >= a certain threshold value)

Dear Dimitri, thank you very much for your kind and prompt advise!

I twisted your suggestion a little bit and I got the following results:

The Tableau Public file is shared :

http://public.tableausoftware.com/views/histogram_04/Sheet3?:embed=y

The cummulative frequency distribution is created according to :

http://kb.tableausoftware.com/articles/knowledgebase/cumulative-distributions

• 3. Re: [help] how to create chart with X axis: value ranges; Y axis: % of case (value >= a certain threshold value)

I see, you wanted % of each bucket, but only for positive satisfaction. There is always a danger of misreading intent while reverse-engineering the desired output.

I am glad it worked in the end.

• 4. Re: [help] how to create chart with X axis: value ranges; Y axis: % of case (value >= a certain threshold value)

Hi, Dimitri, you advise is clear and I'm to be blamed for not explaining my intention clearly.

Still referring to the same data set, I want to calculate the % of people who is satisfied (satisfaction=1) for each brightness band. (without using the satisfaction groups created this time)

e.g. in the chart shown below which is a frequency distribution for the 8 brightness value bands (5-10, 10-15, ...) for all the cases

I'd like to create a chart with its x axis being the same as the one shown above, but with the y axis being the % of people whose satisfaction value is 1 for each brightness band or bin.

i.e. currently, there are 10 person (4.02%) whose brightness values are <=10, and among these 10 people 4 have a satisfaction value of 1, so the y axis value for the "5" brightness band should be 40% (4/10). Similarly, there are 46 (18.47%) people with a brightness <=15, and among these 46 people there are 25 with satisfaction of 1, so the y axis value for the "10" brightness band should be 54.35% (25/47), etc...

I've played with the table calculation function a bit, but still don't quite get it ...

May I ask how to achieve this?

The base chart is shown here:

http://public.tableausoftware.com/views/histogram_07/Sheet2?:embed=y

The Tableau Public workbook file is attached as well.

Thanks again!

- J

• 5. Re: [help] how to create chart with X axis: value ranges; Y axis: % of case (value >= a certain threshold value)

Joe, we have public holiday on Wed, I will try to take a look after that (unless someone chips in before then).

• 6. Re: [help] how to create chart with X axis: value ranges; Y axis: % of case (value >= a certain threshold value)

Hi, Dimitri, thank you!

I've explored a bit and pls kindly review the charts created and the annotations from chart 01 to 08 explaining the approach:

http://public.tableausoftware.com/views/histogram_08/08?:embed=y

Wish you a great holiday!

- J

• 7. Re: [help] how to create chart with X axis: value ranges; Y axis: % of case (value >= a certain threshold value)

Looks like you figured it out, Joe.

Since I have no idea what I am looking at (I don't understand the data behind the graphs), it is difficult for me to comment, but I think it all worked as you wanted.