I'm currently looking to count the number of instances a values shared across multiple dimensions. For example, say I have the following set of data:
And I want to return something like:
But ideally in the form of a bar graph. I want to keep the names associated with the data, so I can filter lets say by all "Bobs" or all "Hannahs".
Does anyone have any advice on how to do this in Tableau?
Here are a couple of ways you may be able to do this.
1) Create a calculated field for each food type. This is a bit cumbersome and you would need to add new ones for any new foods added. You calculations would look like this:
Hamburgers:
SUM(IF [Food1] = 'Hamburgers' OR [Food2] = 'Hamburgers' THEN 1 END)
Then you would make use of the Measure Names and Measure Values built-in fields.
2) You can normalize your data. If you are referencing a Excel or Text file, you can do this right in Tableau. Simply go to the Data Source tab, select the Food fields, and choose to Pivot them:
Goes to:
Now you can do:
Finally, both results support creating a bar chart:
Related
I used an Apriori algorithm to view the frequent relationships in the dataset and I want to do a dashboard to better visualize this data but I don't know how to do this filter.
This is the bar chart that I created to show the support (amount of times something happend) and the confidence (probability of B happening given A) of these associations:
Apriori Chart
Next to it on the dashboard, I'll have a table with the full dataset used in this Apriori analysis where I have more information such as ID, Income, Hours Worked, etc:
Table from different data source
How can I create this relationship? The two data sources don't have a column in common that I can use for that.
I would need some way to:
Split the values in the antecedents columns by comma and filter only those columns with value equal to 1 in the other dataset
**Dataset A**
'Age Range <=30, Joblevel 1, Maritalstatus Single'
->
'Age Range <=30'
'Joblevel 1'
'Maritalstatus Single'
**Dataset B**
'Age Range <=30' == 1
'Joblevel 1' == 1
'Maritalstatus Single' == 1
Clicking this would filter the table next to it
Is there any way I can do this in Tableau?
You can download the tbwx i used in this example here https://community.tableau.com/servlet/JiveServlet/download/1083124-384949/Apriori.twbx
Thanks in advance for the help!
I am not able to check your twbx on the machine I'm using but I think you should be able to do this. The fields in the 2 data sources need to match so manipulate the data sources the make this happen.
For data source 1 there's a function SPLIT which will mean you are able to split the comma separated string to 3 fields.
Putting those 3 fields to the Detail shelf of your bar chart (or even Rows and hiding the header) will mean you can use them in an action filter.
Your second data source is a cross tab - post pivot. You should be able to pivot this data source. Highlight the measures and pivot them. This will give you the field Pivot Field Names and Pivot Field Values.
You only want to keep those with a value of 1 so create a calculated field
[Lookup1]: IF [Pivot Field Values] = 1 THEN [Pivot Field Names] END
Duplicate this field twice so you have Lookup1, Lookup2 and Lookup 3.
Then you should be able to action filter the table.
In the action filter set it up so SplitField1 = Lookup1, SplitField2 = Lookup2, etc.
Fingers crossed this works, I haven't been able to test so I am pulling it out of my head.
Is it possible to create a line chart using summarized data from dataset?
My scenario is the following:
detail section: each one of the cells are one output field in the dataset
JAN FEB MAR
item1 R X R
item2 X A R
item3 R R R
footer section: here we count the occurrence of each value per month by using count aggregation elements and filter by value
TOTALS:
R 2 1 3
X 1 1 0
A 0 1 0
And what I need to do is to add a chart that shows something like this:
Needed Chart
And what I have is this:
report output
report design
In EXCEL this is really simple, but I cannot figure out how it can be done in BIRT.
I thought something like to create a new data set with 3 output fields for R, X, and A and each row will be one month, so I will have a transposed table and that way it will me much easier for graph it. But I cannot do it using the aggregation fields, and I cannot find out how to it with the output fields from the main dataset.
Any ideas? If you need the source rptdesign file I can provide it to you, but the logic in it could be hard to understand.
Any help is appreciated, and thanks in advance.
Have a great 2020.
First of all, a BIRT rule of thumb: if you need aggregations outside of a table, create them outside of a table. Do not try to access values in a table from the outside. It is possible, it may sometimes be the only solution, but it usually messes up your whole report, it is hard to debug, and even harder to maintain.
Aggregate
As your dataset looks quite simple and you already know how and what to aggregate, your first call should be computed columns in the dataset:
Here you aggregate in the language according to your datasource. If that is SQL, I guess a COUNT and GROUP BY statement will do the job.
Create all the columns you need for your graph here.
BTW: Computed columns are usually the silver bullet in BIRT. I use them for almost any pre-computation or custom field creation.
Visualize
You did not mention the library you are using for graphs, so I will assume you want to use the basic BIRT graphs. The basic charts with the months on the x-axis will do your job. I just want to add here that you have two options for multiple series:
You can either prepare your dataset so that you can feed the graph with a series per type (one line in your chart example) OR maybe easier: use optional y series grouping on your computed columns (as mentioned):
This way your graph will create the separate series for you. I hope this helps!
If you get stuck with the basic birt grpahs in general, you might want to think about finding a JavaScript-based graph library that does exactly what you need and implement that. Remember: you can put almost anything that is based on JavaScript into BIRT.
Final remark: For the sake of your report end users, please use a multi-bar chart. Line charts are not readable for overlapping values.
Thanks Kekzpanda for your help and time helping me in this question I had.
After struggling for a while I finally reach for a solution transposing the "table" of aggregations I had in the table footer, by using javascript arrays and an extra dataset and here are the steps I did in case someone else have the same problem:
For example, you need to transpose a table of 10 columns and 3 records
In report initialize method create an array with [10,3] dimension
// bi dimensional array indexes
var i=0;
var j=0;
// array definition and initialize it using 'for' iteration
var matriz = new Array(10);
for (i=0;i<10;i++){
matriz[i] = new Array(3);
}
// restart the array indexes in case you need to go through it in the future
i=0;
j=0;
Then you need to save the aggregation field value in one of the positions in the array. For that click on the aggregation field and go to the onCreate method and add the following code:
matriz[0][1]=this.getValue();
When finish all the aggregation fields, you will have an array with the transpose table. CHECKED!
Move the data in the array to the new dataset, select the fetch method and add something like this:
if(i >= array.length) return(false); // when you finish going through each item in the array.
row["A"]=matriz[i][0];
row["B"]=matriz[i][1];
row["C"]=matriz[i][2];
i++; // increment the first index by 1 to move to next row in the array
return (true);
Now you have your new data set with the transpose data.
Now work with this data set and graph the data, create the different series in the graph design for each column in the dataset.
Hope this help.
Bye.
How can I extract the IN count portion of a Tableau set? I can see the IN/OUT counts when I drop the set into Text but can't figure out how to get at the IN value by itself.
Ultimately, I want to create a Pie Chart of three sets with just the IN counts as the measures.
I am using Tableau Public if that is a factor.
You have to be a little careful about specifying what you wish to count.
One way to think of a set is as a Boolean function that gives a value to each data record denoting whether that record is associated with the set.
Another way to think of a set is as a mathematical set whose members are a subset of the values for some discrete field. (Or Tuple of fields)
The difference between the two views is really just a mindset, whether you consider the set as a Boolean function whose domain is a data row in the data source, or whose domain is the field on which the set definition is based.
Say you are looking at Tableau’s Superstore data set where each data record is a line item for a product attached to an order.
If your set is based on the field Region, say its called [My Favorite Regions] and currently contains {“East”, “Central”} do you want your count to be 2 (i.e. the number of regions in the set) ? Or do you want your count to be in the tens of thousands (i.e the number of line items on orders from the regions in the set)? Or something in between, maybe the number of distinct orders (i.e. order ids) within the selected regions...
If you want to count data rows that are associated with the set, you can simply filter by the set and calculate SUM([Number of Records[). If you want to count the regions in the set even though the level of detail of the data is at the order line item level,then you’ll have to use either a COUNTD to count the distinct regions, or some approach to specify what it is you want Tableau to count.
For example, put your set on the filter shelf, and show COUNTD(Region) which could be slow for very large data sets. To get the same effect without an explicit filter, you can define a LOD calculation such as:
{ COUNTD(if [My Favorite Regions] then [Region] end) }
Or you could use a table calc with the SIZE() function to do the calculation in the Tableau client instead of by the data source.
Not sure what your data looks like but you could set a certain condition when creating a set or split the IN/OUT into two different sets.
Here's a link to sets in Tableau.
You can do this with an if statement
IF [set] = TRUE THEN 1 ELSE 0 END
Then I suppose you could sum this calculated field
The most common usage is when you have a lot of categories and want to create an 'Other' category based on the categories that aren't in a set, if the set is a "Top N Set"
To do this:
IF [set] = TRUE THEN [dimension] ELSE 'Others' END
I am trying to use a calculated measure as a way to filter my data, but it's looking more difficult than expected. Let me explain through an example.
I have data of the following type, with two dimensions - one is a unique ID, the other a category - and four measures.
Initial table
My first step is to rank each element by its score, where the ranking is evaluated within the same category. I therefore create a new measure:
=aggr(rank(sum(Score1)), Category, UniqueID)
I do this for all three scores, resulting in three new calculated measures. My final calculated measure is the average of the three rankings. Below the example, the calculated measure of interest is the one in bold. Note that in my real world calculation I directly evaluate 'New Measure', without creating the intermediate columns 'RankingScore'.
Data with newly calculated measure
Note that this measure is tricky, as it changes according to previous selections. Say, for instance, that I select only entries with 'Amount' > 1000. The relative rankings will change and therefore also 'New Measure'.
In my actual App I need to filter my entries by 'New Measure', after I've done some previous selections on fields like 'Amount'. If it simply were a field, I would normally have created a filter pane, our used the qsVariable extension to have a slide range, to select only rows with 'New Measure' above a set threshold. Unfortunately it seems I cannot do that with my calculated measure.
How would you approach the problem? I was wondering, for example, if it were possible to 'convert' my new measure to an actual field, after all previous selections have been done, but perhaps this is nonsense.
Thank you in advance, and apologies for the long post!
If I'm understanding correctly, I believe this solution should work:
Create a variable for your slider: new_measure_slider.
Create a New Sheet Object -> Slider/Calendar Object.
Configure your slider to control your new new_measure_slider variable.
Create a calculated dimension in your chart substituting your 'New Measure' formula (the one you stated was an average of the three ranks). It should be a conditional like this:
=if(aggr([your average formula here], Category, UniqueID) >= new_measure_slider, [Category], null()).
Basically, compare your formula to the new_measure_slider variable. If true, use the Category (or UniqueID, whichever you need) as the dimension, if false, null().
Check the 'Suppress When Value is Null' checkbox on your new dimension. This is key. This is what will actually filter your chart.
In the chart properties, Presentation tab, click on your new calculated dimension and hit 'Hide Column'. We don't need to see this because we are using it only as a filter.
You can tell QV to ignore your filtering in the field Amount by adding "Amount=" to your set analysis.
I dont know how your average calculation looks like but maybe:
(aggr(rank(sum({<Amount=>} Score1)), Category, UniqueID) +
aggr(rank(sum({<Amount=>} Score2)), Category, UniqueID) +
aggr(rank(sum({<Amount=>} Score3)), Category, UniqueID)) / 3
I'm using Tableau Desktop 9.0 on OSX. I have data (loaded from a local CSV file) that looks like this:
code,org,items
0212000AA,142,10
0212000AA,143,15
0313000AA,142,90
0314000AA,143,85
I want a chart that shows the number of items beginning with 0212 as a percentage of all items, for each organisation. (I mean as a percentage of the organisation's items - for example, in the above, I would like to show 0.1 (10/(10+90)) for organisation 142.)
I have been able to get part way there, by adding org to Columns, and SUM(items) to Rows. Then by adding a Wildcard filter on code, for starts with 0212.
This shows me the number of items starting with 0212, by organisation.
But what I don't know how to do is show this divided by the value of all items for the organisation.
Is this possible in Tableau, or do I need to pre-calculate it before loading my data source?
One way is to define a calculated field called matches_code_prefix as:
left(code, 4) = "0212"
You can also define a parameter called, say, code_prefix to avoid hard coding the prefix string:
left(code, 4) = code_prefix
And then show the parameter control for code_prefix to allow the user to interact with it.
If you use this new field as a dimension to separate SUM(items) according to those that match the prefix and those that don't, you can then use a quick table calculation to get the percent of total.
For example, you can place org on the Rows shelf and matches_code_prefix on the Columns shelf, and SUM(items) on the Text shelf to make a table. Then under the analysis menu, turn on grand totals for both rows and columns to see the behavior. Next, right click on SUM(items) and choose Quick Table Calc->Percent of Total. Tableau will display the percents of total in the table.
If you want the percent of total defined differently than the default, then right click on the measure again and set Compute Using to a different value such as matches_code_prefix in your case. It's usually better to set compute using to a specific field.
If you only want to display the value for the matching case, select the column header you don't want to see and choose hide. You can also turn off the grand totals from the analysis menu when you are done.
When you are confident in the values in your table, you can turn it into a bar chart for example by moving matches_code_prefix to the detail shelf and the measure to the Columns shelf.
--
The above is the drag and drop approach. If you prefer to hard code everything in a single calculated field that is calculated on the database side, you could instead define a calculation such as:
zn(sum(if matches_code_prefix then items end)) / sum(items)
Then set the default number format for that field to display as a percentage