trelli/multiple scatter plot and pie chart

[JimmyCruyp]

Hello guys, could we create trelli pie chart? And if not, is that possible to have the code to create this with Deneb module?

I have the same question for multiple Scatter plot .

Thanks a lot,

Jim

[giammariam]

Hey @JimmyCruyp. I implemented some changes and have a version working in Power BI.

You want to make sure that the Summarization for the Result field is set to Don't Summarize (see the below screenshot).

Note - you'll need to play with the radius percentages and the spec width and height based on the size of your visual (see below screenshot)

You can download a working .pbix here

[JimmyCruyp]

It's awesome. But small question: if the users are not report creators and will not be able or will not have the authorizations to modify the code in Deneb, that means that the graph risks having a bad format? It does not automatically adapt in format according to the number of sector graphs we have? It is very disturbing though....

[giammariam · Accepted solution]

I coded the facets to dynamically handle the row/column layout, so adding/removing charts within reason should not be an issue.

 

Regarding my comment about adjusting those width/height/radius percentage values in deneb, I meant that one would only need to do that if they modify the visual's width and height in the formatting properties (see below screenshot). This is a limitation of faceting in vega-lite; currently the facet cell size cannot dynamically change based on the size of the canvas (yet). I cannot think of a scenario where a user could adjust the visual dimensions but would not have access to being able to update the spec in Deneb...

 

[JimmyCruyp]

You are a monster. Thanks a lot. Now I will need the same but for scatter chart. I will create a new post. Thanks superman

[giammariam]

Hey Jim, this could certainly be done with Deneb. It could also be done with the native matrix and DAX generated SVGs (similar to your correlation matrix post).

The latter would be a bit tougher to pull off, so I'd recommend going with Deneb this time. If you provide some data, I'd be happy to crank out a vega-lite spec. 

[JimmyCruyp]

Type	Voice	Result
A	Power	10
A	Then	5
A	Zeld	32
A	Link	8
A	Nintend	2
A	Sega	7
B	Power	8
B	Then	4
B	Zeld	6
B	Link	8
B	Nintend	20
B	Sega	52
C	Power	9
C	Then	13
C	Zeld	25
C	Link	2
C	Nintend	2
C	Sega	3
D	Power	10
D	Then	20
D	Zeld	54
D	Link	10
D	Nintend	30
D	Sega	10
E	Power	3
E	Then	65
E	Zeld	9
E	Link	24
E	Nintend	35
E	Sega	22
F	Power	70
F	Then	17
F	Zeld	21
F	Link	80
F	Nintend	7
F	Sega	10
G	Power	9
G	Then	90
G	Zeld	20
G	Link	20
G	Nintend	100
G	Sega	20
H	Power	5
H	Then	5
H	Zeld	120
H	Link	5
H	Nintend	5
H	Sega	75
I	Power	33
I	Then	130
I	Zeld	33
I	Link	1
I	Nintend	0
I	Sega	0
J	Power	40
J	Then	11
J	Zeld	22
J	Link	11
J	Nintend	50
J	Sega	24
K	Power	3
K	Then	7
K	Zeld	60
K	Link	21
K	Nintend	12
K	Sega	22
L	Power	70
L	Then	10
L	Zeld	0
L	Link	10
L	Nintend	20
L	Sega	80
M	Power	30
M	Then	22
M	Zeld	25
M	Link	1
M	Nintend	140
M	Sega	19
N	Power	22
N	Then	27
N	Zeld	30
N	Link	5
N	Nintend	30
N	Sega	6
O	Power	22
O	Then	33
O	Zeld	8
O	Link	9
O	Nintend	38
O	Sega	31

[giammariam]

@JimmyCruyp, while I personally don't love pie charts, I do like your choice of dummy data 😀.

There's a lot that can be configured and formatted here, but I kept it simple. Let me know if you need any tweaks.

You should be able to copy the spec into Deneb. You'll want to drag in the Type, Voice, and Result fields into the data pane, and remove the url and format properties under the data property in the spec.

If this gets you going, please consider liking this reply and choosing it as the solution so others can find it. If you need additional help, just let me know. 

Here's the gist

spec: 

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "name": "dataset",
    "url": "https://raw.githubusercontent.com/Giammaria/PublicFiles/master/pbi/pbi_community_help/trellis_pie/data/data.tsv",
    "format": {"parse": {"Result": "number"}}
  },
  "transform": [
    {"joinaggregate": [{"op": "distinct", "field": "Type", "as": "type_count"}]},
    {"joinaggregate": [{"op": "sum", "field": "Result", "as": "result_total"}], "groupby": ["Type"]},
    {"calculate": "toString(datum['Result'])+' out of '+toString(datum['result_total'])", "as": "result_out_of"},
    {"calculate": "format(datum['Result']/datum['result_total'],'.0%')", "as": "percentage"}
  ],
  "mark": {"type": "arc"},
  "encoding": {
    "facet": {
      "field": "Type",
      "type": "ordinal",
      "columns": 5,
      "sort": {"field": "Type"}
    },
    "theta": {"field": "Result", "type": "quantitative"},
    "color": {"field": "Voice", "type": "nominal"},
    "tooltip": [
      {"title": "Voice", "field": "Voice"},
      {"title": "Result", "field": "result_out_of"},
      {"title": "Result %", "field": "percentage"}
    ]
  }
}

[JimmyCruyp]

Hi,

Is that possible to have also a bar to set the graphs dimension? As you have done for the multiple treilli? This could be great!!

Jim

[giammariam]

Please make a new post and tag me with this request. I'd be happy to.

[JimmyCruyp]

thanks a lot this looks great. But i dont see anything appearing in Power BI. Maybe because i dont have any configuration? Could you please provide me also the config? Sorry i am very bad at this.

I have added the 3 fields and remoove the url to obtain something starting with this : {
"data": {"name": "dataset"},
"transform":

Could you also please add a percentage in each pie per voice (legend)? fi