Transform vs Modelling
The Transform capability is great, because i can do everything I need with the UI. This is not the case with Modelling.

Modelling
Having used PowerBI Desktop on many datasets, I found that it should be made simpler for non DAX experts.A UI based interface including some "AI" Assisted intelligence could easily achieve the below and make the product reach a new category of users.
1. Matrix and Table visuals
• Key features such as Column renaming, Cumulative values, Comparisons Year on Year etc. are missing and require quite a bit of DAX fiddling. I know that you are working on a new visual.
• they optionally output a virtual table: it is much simpler to generate the data required visually that to fiddle with DAX (again, with code showed to user, this would be a fantastic training tool). They could serve as a model to train users or be reused.
• Change the source table(s) name(s) : often you want to "test" on a table and reuse on another one
2. Calculated Tables
• Adding columns can be done in Modelling, if it were not there, ADDCOLUMNS would need to be used
• Appending several identical tables together : need a UNION , this could be UI based
• A significant amount of time is required to create, prepare "Utility tables" such as Code to Literal, standardising literals, grouping literals, calculating rankings (this being quite difficult) : an "Assisted Intelligence" UI would make life much easier
3. Ranking, Top n
• This is quite complex (despite the Top N filter) when you want to show the Top n as distinct values and the rest grouped as other. Especially when this is not related to time, but to a category etc
4. Where are the modelling instructions stored?
5. Common formatting options
• All my numbers are formatted with a Thousand separator and no decimals. Why can't I specify it once?
6. Utility tables
• There are always many "Utility" tables such as Code to Literal value. Some are common across PBIX files (i.e. County code to County). I could not figure out how to Reuse them as opposed to Re-create them in each file

Power Bi just implemented custom themes but what this does is applies the same theme to all the visualizations. This is could be very confusing to the end user. Suppose you have multiple metrics (like sales, profits, margins, etc.) that are sliced by geographical region as one dimension and say Item category as another dimension. Currently the same top 5 colors of the theme would be used for the top 5 geographical regions and top 5 item categories.

Each dimension should be using a distinct color palette. That is say California and Clothes should not be using the exact same color. The both could be blue in color but they should be distinguishable shades of blue.

Currently this has to be done manually. Even MS Excel has multiple color palettes for use on different charts.

A similar issue is also seen in hierarchy colors. That is the state of California and city of Houston (drilled-down from Texas) could have the same color.

For different dimension values there should always be a different color. And there should be color palettes/themes pre-built (as doing this manually is time-consuming and not intuitive).

I have 5 reports with 5 tabs each. I also have 5 dashboards, each linked to one of the reports. All of these originate from a single PBI-desktop file with all 25 tabs in them.

Each time I make some changes in the desktop file and publish it, I have to save the main report into the 5 copies and remove 20 out of 25 tabs in each. Then I have to save the tabs to the dashboards again. This is a tedious, time consuming proces. I wish this proces was somehow automated or improved.

Any of the following would help immensly:

- Link the tabs from the main PBI report to the online tabs spread over multiple reports. Maybe by an id-number. They all use the same dataset anyhow. On publish, these should automaticly overwrite the online tabs with the same id.
- A fast way to remove tabs without having to the report. Currently, clicking on the cross on each at the bottom goes pretty slow and takes a few seconds of loading time.

Here I have a query that populates result sets dynamically by taking an input parameter.Now the result sets does not have equal number of columns neither they have same column name.So , while using that query to visualize a result set in a table,when I pass another value for the parameter in the query,it creates different result set and as the columns of the new result set is not included in the table value fields,the table view gets invalid. And I have to show result of the query in same visualization.

Is it possible to have a visualization on Report interface that acts just like what Data viewer pane does,i.e., it always shows output of the query whatever the result set is,like a flexible table that adapts to it automatically,or like the Data viewer pane.