Counting document frequency of keywords

[Tommy_Monk]

I'm performing some statistical analysis on keywords and have been stumped trying to query for count of occurrences.

Here's a simplified example of what I wish to achieve;

Example Documents

{ "Colour": "Red" }
{ "Colour": "Green" }
{ "Colour": "Blue" }
{ "Colour": "Red" }
{ "Colour": "Red" }

Output Description

Count of documents that have Colour only occuring in 1 document
Count of documents that have Colour only occuring in 2 document
...
Count of documents that have Colour only occuring in 50+ document

Output Example

[
   {
      "Key": "1",
      "Count": 2 // Green and Blue only occur in 1 document
   },
   {
      "Key": "2",
      "Count": 0
   },
   {
      "Key": "3",
      "Count": 1 // Red occurs in 3 documents
   }
]

I would like the result to be constrained by a query, I've considered using a Terms agg with partitions and counting in code but this seems sub-optimal.

[Tommy_Monk]

Is there a more appropriate place that I should have asked a query related question like this ?