DavidChurchman Coach

Sorry, the y-axis is actually this; sum(ifnull(Promoter,0))/count(DISTINCT Response ID)-sum(ifnull(Detractor,0))/count(DISTINCT Response ID)

I removed the sort, and the y-axis is COUNT(DISTINCT Response ID)

To generalize @MarkSnodgrass 's solution a bit, you can use most Unicode characters to hack different visuals like this into your tables, summary numbers, chart labels. There a lot of lists of available Unicode characters, but this is my go-to: https://www.vertex42.com/ExcelTips/unicode-symbols.html

You could create a rank column using "Rank&Window" in MagicETL. Here's a dataset where I've ranked 4 products by month, then created a beastmode for "Top 3" or "Everyone Else": Here it is filtered for Top 3: In a dashboard setting, you could create a slicer with that BeastMode, and filter the slicer for just "Top 20", so…

I would suggest a re-structure of the data to something like this: Then you could do a waterfall: Or a running total bar, which sounds more literally what you're asking for: Here's the MagicETL I used to restructure the data:

Super helpful, @RobB !

You can use "color rules" in a lat/long map. It doesn't appear in the "chart properties" panel they way it usually does, but you can open them using the button on top, then apply rules to the variable you chose for your "series for color".

You can create multiple color rules, not just one. If you use a BeastMode to label your "New" and "Existing" values, use that for your "series for color", open up the color rules, you should be able to create a color rule for "New" and a color rule for "Existing". Alternatively, you can create a color rule for the whole…

From what I can see in your flat file, you have 4 rows, 3 which have a SourceCode=AP, and one with a "Blank" source code. There is no TrasnactionAmt for that Blank SourceCode, so it is returning a null, which you convert to 0 with your ifnull. If you want the ELSE to return the total of 1,015.56, then remove the case…

You could create a new column that concatenates "Section" and "State" and filter based on that, using the "Fixed" calculations to calculate percent increase. CONCAT(Section, State). Then you can filter on "A-Texas", "B-North Carolina", and "C-Kansas".

This appears to be a graph that is filtered on "price differences", "valid" and "invalid", so I'm not sure what the problem is. Maybe it would help to describe or sketch what you were hoping to see after your filter?

Yes! The Flex Table has potential to be the most useful card type, making big sparklines-style tables, but the options are so limited, I can't use it for anything.

Yes, more table formatting options, please!

Did you try skipping the transform to null? Case when coalesce(trim(`commitment`), '') ='' then `loan` else `commitment` end I can't find this written anywhere, so maybe I'm making it up, but I'm pretty sure you can't THEN NULL to return nulls. I'm surprised it's not throwing a syntax error.

It might help if you flesh out your example a bit more. Based on what I understand, it sounds like you could create two table cards for your dashboard and two filter cards, and then you could edit the interactions on the filter cards using the "change interactions" menu to select which filter applies to which table card.…

You could use Origin as the 'Series for Color' in the Lat/Long map. This would tie the origins to their destinations, so when you hover on an origin, all the destinations would appear, and you could click to filter by that series. You'd want to restructure your data a bit. In a magicETL, you could select all of your origin…

Michelle's suggestion is good, but the bullet chart is designed for this type of situation. You could create BeastModes for each range, per Michelle's suggestion, and then use them as the different target ranges on your card (range 1, range 2, etc.). "Worked Hours" would be the same color throughout the chart, but would…

Is there any kind of grouping that would be useful? I find with large tables, it's helpful to start with a grouping variable and check off "show sub-total rows", which allows you to collapse the groups and help them navigate more quickly around the table by expanding/collapsing those groups. My understanding of 'dynamic…

Similar to what you're saying, you could use an ETL to restructure your data to something like: Type Index | Description | Month | Actual/Budget/Variance | Amount To do that, you could: split your dataset using two filter tiles into actual and budget Join them back together on Type/description/month Use a formula tile to…

The forecasting chart has fields where you can plug in lower/upper bounds of forecasts. KB article on forecasting chart

You can add multiple conditions to a filter tile, then there is the option of "all" or "any" of the rules to be true. Basically, logical "AND" or logical "OR". If it's more complex than that, you can create a "formula rule" within the filter tile. KB article about filter formula

Depending on the setup of the card/data, I think you should be able to change the Beastmode to just COUNT(DISTINCT EMPLOYEE_ID), add it as a Tooltip so that it's at the right level of aggregation and then use it as a filter.

A possible approach: what if you appended department-level copies of the dataset to the bottom of the dataset? Then you could give departments access to their rows via PDP and the default filter could be for single-select filter for "View my department only" or "All data". You would just have to be careful that if they…

The Total is your "Hover Legend". When you're not hovering, it shows the total, when you are, it gives details around the slice you're hovering over. You can choose the option to hide the hover legend to remove it. The reason it's showing in the graphs on the right but not the left is they are sized a little smaller. A…

That's a bit different. It's not a simple sum, but the sum of the max value by user. That would require a fixed function (https://domo-support.domo.com/s/article/4408174643607?language=en_US) that could look like this: sum(max(number) fixed(by id)) You could still either make 3 cards or create a variable that allows a user…

Maybe you could create a filter on a BeastMode like this, assuming you don't have negatives that would make it inadvertently add up to 0: CASE WHEN SUM([All those value columns]) FIXED(BY [All those row columns]) = 0 THEN 'Exclude' ELSE 'Include END You could wrap all your value columns in an absolute value or square them…

I frequently encounter problems when filtering/sorting columns with null values. Marks' suggestion is good, because it will give you a column with no nulls.

In addition to Mark's suggestion of turning on the 'Show Current Date' line, you can also calculate your percent complete to always be up to today's date, like: case when DATEDIFF(CURRENT_DATE(),Start Date) / DATEDIFF(End Date,Start Date) > 1 then 1 else DATEDIFF(CURRENT_DATE(),Start Date) / DATEDIFF(End Date,Start Date)…

Not totally the same, but you can do this with a filter view.

Assuming by row, you mean everything that will display as a row in the pivot table and not a row from your original dataset. If that assumption is correct, then the 0s will be all the values associated with what defines your rows and columns With that in mind, the BeastMode GrantSmith mentioned above would need to be a…