Vincent Thuilot

Hello, Could this be a solution? Here what i did: 1- Calculated the Lot number and formatted it as Integer (i assume this logic would need to evolve if the 2 characters change) 2- Unpivoted as suggested by Olivier 3- Calculated a new column to isolate the value corresponding of the last 6 lots case WHEN [ID]>(UniqueCount([ID]) - 7) then [Value] END Again, you will surely have to adapt it based on your lot coding logic 4- Create a Combination Chart with a) the Value of the newly calculated column and b) an Average line with all Values and of course here the avg formula that i used: Avg([Value]) OVER (All([Axis.X])) 5- Filter based on the new column, i put greater than 0 for the example but you could set it to Not Null or something else that better matches your needs. 6- Add the Treillis, adapt Y axis scale for each treillis panel and voila :) Not sure it will satisfy your needs but maybe it will help in your thinking. Vincent Combination Chart Try.dxp

Wonderful, it works like a charm! Many thanks to you and your colleague Rae Chen, it was not that straightforward! I will use it right away for the dashboard I was working on and will use it as a reference for future calculations like these. Very instructive to see all of these expressions working together.

Hi David, I attached an example of what i need, please see DXP attached. we are on 11.4 version. As mentioned earlier, i am not interested in showing the CAGR (Compound Annual Growth Rate) or a basic Sales Growth. What I am interested in is showing the evolution of the market share that would show totally different numbers (Sales Growth can increase whilst MS can decrease if some other clients were performing even more on a given year). I would like to keep it into a cross-table because it's very dynamic thanks for the support, please let me know if you need anything else. Vincent YoY %MS Growth.dxp

Hello, Assuming you don't have too many bullets another option could be to create your table in excel, do your initial formatting there. Then, copy paste the formatted table in a text area and insert Icon in the needed cells. It's not easily scalable but it will do the job if you have a fixed number of KPIs Vincent

Hello, In a cross-table, i would like to compute % market share evolution year over year for clients appearing on the row axis . Calculating market share on a given year is rather straightforward: Sum([Sales]) THEN [Value] / Sum([Value]) OVER (All([Axis.Rows])), where the rows are the different clients (just one dimension). But how can I calculate the market share difference Year over Year for each client and summarize it fine in a cross table? Example: - if client A has a market share of 17% in 2023 and 17.3% in 2024, I want to show +0.3% in 2024 within the cross table - if client B has a market share of 16.2% in 2023 and 15.3% in 2024, I want to show -0.9% in 2024 within the cross table I tried several times but got limited as i couldn't leverage the THEN [Value] command several times. NB: I am not interested in the YoY sales growth, that is also pretty straightforward: Sum([Sales]) THEN ([Value] / First([Value]) OVER (NavigatePeriod([Axis.Columns],"Year",-1))) - 1. I have the feeling that I missed something obvious! Thanks in advance for your support! Vincent

Hello, what about something like this Assuming that you are fine with adding RowId, you can compute the lowest rowId by subcategory like this: CASE WHEN [RowId()]=Min([RowId()]) OVER (INTERSECT([SubCategory])) THEN [Value] ELSE "##" END and then use FIRST in the cross-table, after sorting by RowId it should work out fine. You can later hide the RowId column with an IronPython Script if this is necessary: something like this: # 'vis' is a script parameter # Import Modules from Spotfire.Dxp.Application.Visuals import * # Set row header column width to 0 (i.e. first column) vis.As[CrossTablePlot]().RowHeaderWidths.Item[0]=0 I attach the dxp i quickly worked on, hope it helps! Data - Sheet1.dxp

Hello, Quick question on this topic, do you plan to have Certification Days as we had in the past? It was really nice to push our organization to get further into Spotfire and take an exam to test our skills. Once or twice per year would be awesome :) Best regards Vincent

OK thanks for the suggestions Olivier and Barchiel. I have submitted 3 ideas to maybe have something out-of-the-box that would make it easier to end-users. Each of these are complementary. I believe that enhancing a little bit the design options could really render some dashboards MUCH better :) - Visual Transparency https://ideas.spotfire.com/ideas/TS-I-9697 This would allow users to have a better grip for each visual's transparency on a given page (not just 0% or 100%) - Background Gradient Option: https://ideas.spotfire.com/ideas/TS-I-9698 This would allow to users to really enhance the background of the Canvas or of the Visualizations - Multiple Canvas Styling https://ideas.spotfire.com/ideas/TS-I-9699 Today when we edit the Canvas Styling options, it is being applied on all pages on all visuals at once. Having the possibility to apply a specific canvas on specific page(s) and another one to other pages could provide more flexibility. -I also came across this: Single Visual Background: https://ideas.spotfire.com/ideas/SPF-I-4906 This one already got 72 votes and really shows the need. Hope it helps! Vincent

Also, it's the page background that must be a gradient, the visual background would simply be transparent. Hope it's clear :)

Hi Olivier, thanks for your message but i am looking for the gradient option here. I'm afraid it's not available in what you proposed.

Hello I would like to replicate a visual like the one attached (see that there is a radial gradient background) in Spotfire but it seems there is no option to control visual background transparency level. As a result, all my charts have a white background or a single color background if i use dark mode. Is there maybe a workaround? This could enhance the overall design of a dashboard

Hello Gaia, thanks a lot for sharing, this is absolutely great work! You totally understood the logic of it and was able to make it real. The main difference with Tableau is that in Tableau this is managed through an out-of-the-box menu that makes it quite straightforward and flexible. Interestingly, each created set there is automatically available in the data pane and visible as a "set" so one can see it was created on top of the existing dataset. In addition to this, one can also combine two sets when meaningful, this can be quite powerful. Spotfire would benefit so much from having something similar to perform analyses, where sets could be created in a very efficient way without having to do much programming. Please don't get me wrong, the solution you provide is a already good workaround. I just wanted to share that there is a great opportunity to enhance analytical capabilities here. Thanks again for the awesome support! Vincent

Thank you Gaia, you got the point, however in your example it's still a very manual process. How to make this more dynamic: - if a filter is applied - if in your example we don't look at row count but other metrics like sales, profit, etc...and get a visual dynamically updated. Does this make sense to you? Best Vincent

HI Gaia, Indeed, but that would be quite manual work and not super user friendly, for instance how to select the top ten manually from a bar chart like this? Not easy, right? Even when sorted out, it's difficult. Maybe there are no workarounds today. Some of our services are particularly asking for this kind of stuff and I also face these situations at time when having a dynamic behavior to create sets would enhance the interactivity.

Hi Gaia In the screenshots (i forgot to annotate them, sorry!), what you can see is the current result when nothing is done: we have lots of colors and can't regroup the small entities as Other. I'm attaching what we need to achieve, where we could only see the top N (here top 4) and the rest is seen as Other (grey). As you can see, the charts are more digestible compared to what i shared earlier. To perform this is went through the steps i detailed above but it's not really dynamic (sensitive to filtering) and it needs 2 calculated columns. Suppose i need to break-down the Top N and Others in X other categories, i would need to create 2*X columns and X charts. Quite cumbersome I think. Tableau is achieving this through what they refer as dynamic sets (https://help.tableau.com/current/pro/desktop/en-us/sortgroup_sets_create.htm). Hope it helps to better understand. Please let me know if anything is still unclear. Vincent