Supposing I have many plots on a y-x axes. I want to calculate the average of all the curves/points above the “average curve” and the average of all the curves below the “average curve”.
I have created a measure that is the “average curve” using the following calculation:
AverageLF = AVG([LF])
I am trying to create a measure that would express the average of all the points above average using something some sort of conditional averaging like this:
AboveAverage = AVG ( IF [LF]>[AverageLF] THEN [LF] END )
Apparently I get the: "Cannot mix aggregate and non-aggregate arguments" calculation error.
I want to do the same with the median. (median of all the points above, median of all the points below the median)
I am relative new to TABLEAU and this is very important for my work, could you please advise?
Thanks
Correct me if I am wrong,
you want to check if the values in [LF] are > [AverageLF] if yes then take them aside and do an average of those values. is that correct?
If yes then do this,
AboveAverage = IF [LF]>[AverageLF] THEN [LF] END
Now you have a value AboveAverage which has list of all values > [AverageLF]
Now Create a calculated field,
Final = Avg(AboveAverage)
This should have average of all values AboveAverage.
Let me know if this works.
Related
I have a column vector "distances", and I want to select a value randomly from this vector such that smaller values have a higher probability of being selected. So far I am using the following, where "possible_cells" is the randomly selected value:
w=(fliplr(1:numel(distances)))/100
possible_cells=randsample((sort(distances)),1,true,w)
Basically, I flipped the distance vector to create probabilities of selection "w" (if I am understanding randsample correctly), so that the smallest value has the probability of being selected equal to the highest value. To check how well this works, I randomly drew 50 values and by using a histogram, I see that the values are higher than I would expect. Does anyone have any idea on how else to do what I described above?
0 Comments
How about something like this?
let's start with 10 sample distances with lengths no greater than 20 just to demonstrate:
d = randi(20,10,1);
Next, since we want smaller values to be more likely, let's take the reciprocal of those distances:
d_rec = 1./d;
Now, let's normalize so we can create a distribution from which to select our distance:
d_rec_norm = d_rec ./ sum(d_rec);
This new variable reflects the probability with which to select each given distance. Now comes a little trick... we choose the distance like this:
d_i = find(rand < cumsum(d_rec_norm),1);
This will give us the index of our chosen distance. The logic behind this is that when cumulatively summing the normalized values associated with each distance (d_rec_norm) we create "bins" whose widths are proportional to the likelihood of selecting each distance. All that is left is to pick a random number between 0 and 1 (rand) and see which "bin" it falls in.
I'm a new poster here, so let me know if this is unclear and I can try to improve my explanation.
I am trying to find a max value of a curve fitted plot for a certain region in this plot. I have a 4th order fit, and when i use max(x), the ans for this is an extrapolated value, while I am actually looking fot the max value of the 'bump' in my data.
So question, how do I select the max for only a certain region in the data while using a cfit? Or how do I exclude a part of the fit?
LF = pol4Fit(L,F);
Coefs= coeffvalues(LF);
This code does only give the optimum (the max value) of the real points:
L_opt = feval(LF,L);
[F_opt,Num_Length]= max (L_opt);
Opt_Length= L(Num_Length);
So now I was trying something like: y=max(LF(F)), but this is not specific to select a region.
Try to only evaluate the region you are interested in.
For instance, let's say the specific region is a vector named S.
You can simply rewrite your code like below:
L_opt = feval(LF,S);
Use the specific domain region S instead of the whole domain L and it only evaluates the region you are concerned with. Then using max function should work properly for you.
I'm having an issue with tables calculations in tableau. I've got sales per manager and the target. I wanna have a calculation of the targets achievement.
That's what my data looks like and the percentage I wanna get:
Unfortunately, Tableau sums up the percentage and gives me the following results:
Target columns is a calculated field. E.g.
IF [Manager Name] = "Manager 1" THEN "190000"
The % columns is a calculated field and calculated as the following:
FLOAT([Profit]) / FLOAT([Target])
I tried to play with Table Calculations but it never worked.
You are right in that a table calculation would help solve your issue. As you have the % field now, Tableau is just summing the field for each manager to get your subtotal. To get Tableau to calculate the subtotal smartly, as you would expect, you need to modify your % field formula.
Your way: Calculate Percentage for each manager and then SUM for subtotal
Fixed Way: Indicate to Tableau that you want the field to be calculated by summing all profit values first in the partition (Location) and then divide to get the percentage. In this case (330000/532000).
You can modify your % field to look something like the following:
WINDOW_SUM(FLOAT(SUM([Profit])))/WINDOW_SUM(FLOAT(SUM([Target])))
See the below thread for some more information on how the windowing functions can help with aggregate calculations
https://community.tableau.com/thread/200670
I have an excel file which contains 4 columns. First column is time based on seconds and the other three are my function. How is it possible to find a time for specific value in the time column? let me give an example:
let say I want to find where is this value in the second column: 0.7636 located in time column? I found it manually which is located between 6960-7020enter link description here.
So, if I have for example a couple of values, and also considering different functions then it is difficult to do it manually.
I hope to hear your support.
Thanks Sepideh
You have to think about what a good solution is first. Let's call your three functions f1(t), f2(t) and f3(t). Now you have values v=[v1,v2,v3] and you want to know the best matching time value.
What is the best matching time value? You have to find some kind of distance metric, which tells you how close the data matches. As a default, I would use the 2-norm unless you have a reason to use something else. This would be:
%no running code, just a formula
d(t)=sqrt((f1(t)-v1)^2+(f2(t)-v2)^2+(f3(t)-v3)^2)
Now having d defined, you want to minimize it. There are basically two approaches. If you are looking for the closest row in your data, calculate d(t) for each row and take the minimum. Another approach would be to interpolate f1 to f3 so you could fill the gaps between your rows, then again search for the minimum d(t)
You can try something like this, to find position of data in column#2:
data = xlsread('Q1.xlsx');
ref = 0.7636; % your reference value
lb = data(:,2) < ref; % find lower value
ub = data(:,2) > ref; % find greater value
lower_bound = find(data(:,2)==max(data(lb,2))); % find lower value position
upper_bound = find(data(:,2)==min(data(ub,2))); % find greater value position
row = data(sort([lower_bound upper_bound]),1); % find position in column#1
and result will be row = [6960;7020].
I'm facing problem in getting the max % and min % from a table containing % difference values.
Year-----A----------B---------C---------D---------Max %----Max Type----Min %----Min Type
2012
2013---4.30%---4.42%---4.34%---4.38%----4.42%---------B-----------4.30%---------A
The table above shows the % difference in sales from previous year. Thus 2012 shows no % (because there's no 2011). I used table calculation to compute the % difference, i.e. "Percent Difference From", compute using "Table (Down)" and "Previous".
The last four columns are what I'm having trouble doing. I want to get the max % and min % and also the corresponding types. I'm not trying to add the four columns to the existing table, but to get the correct results, as my ultimate goal is to display that results on the dashboard, i.e. on my dashboard, I want to display the highest % and its corresponding type; similarly the lowest % and its corresponding type. For example: on my dashboard, I want to display:
Highest % and type: 4.42% B
Lowest % and type: 4.30% A
So, I need to have the correct formulas to get the max % and min % and their types. These are what I did:
I tried to use WINDOW_MAX and WINDOW_MIN to display the max % and min % on the table but got funky wrong results.
1) I first get the formula in calculating the % difference from the "Customize" button from "Edit Table Calculation" window of SUM([Sales]): (ZN(SUM([Sales])) - LOOKUP(ZN(SUM([Sales])), -1)) / ABS(LOOKUP(ZN(SUM([Sales])), -1))
Then I created a calculated field of the above formula. I named the calculated field "Percent-Diff".
2) I created another calculated filed (named "Max % Difference") using the formula: WINDOW_MAX([Percent-Diff]). But it shows strange results. See image below. I don't know why it gives me 2.78% and 2.91% for 2012 and 2013 respectively. It should be 0% and 4.42% for 2012 and 2013 respectively. Something is not correct.
If it is just SUM([Sales]) instead of % difference, then I get the correct result of showing the max sales using the formula WINDOW_MAX(SUM([Sales])).
3) Also I don't know how to get the corresponding type. I tried using the formula: IF [Max % Difference] = [Percent-Diff] THEN ATTR([Product Type]). But it returns:
NULL
B
I'm not sure if the formula is correct. It looks correct on the result (i.e. "B" is correct), except that it also shows a NULL value which I don't know why. I think it's because I didn't include the ELSE part in my IF formula? But why the NULL value is shown as the first value? I want the formula to return just one value, "B". So, how to only just show "B"?
I've posted twice the problem in tableau forum, but as of now, nobody has answered my problem. I believe that my formulas are incorrect. So, if anyone here can correct the formulas to get the max % and min % from % difference values and also to get the corresponding type, then it'd be very much appreciated. Thanks a million!
It's hard to tell not knowing how your database looks like (as you didn't explicitly presented it, but I can try to infer based on the clues you left on your post). But I could reproduce something like you said using the Sample - Coffee Chain Database, and it worked out well, calculated yoy sales increase by product and then window_max of that.
What you're probably missing is the partitioning. I suggest avoiding using Table or Pane to create the partitions in more complex situations (as it will work only in that specific arrangement of fields), but rather use the dimensions to partition it.
So, your [Percent-Diff] field should be compute using [Date], and your [Max % Difference] should be compute using [Product Type]. IMPORTANT, for [Max % Difference], when you go to Edit Table Calculation, you'll have to choose the Compute using for [Percent-Diff] as well (you can choose on the top of the window)
Your formula to find which type is the max (or min) is also correct (and should only respect the partitions). Nevertheless, it is very hard to have the exact output you're expecting.
What I would do is to create 2 spreadsheets (and later combine them in a dashboard).
The 1st would be what you already got (Each product [Percent-Diff]
The second one I would change your formula (3) to just [Max % Difference] = [Percent-Diff], and use it as filter (filtering only true). I would drag both Date and Product to the sheet (you choose if you want it on columns, rows, or just detail) so I can use them to partition the table. And drag [Max % Difference] to be visualized.
That way you'll only see the product that is the max, and how much is that max.
Hope it helps