Meteor MongoDB aggregate - Collapsing certain values to 'Others' - mongodb

Question

I am working on a toy project with Meteor and MongoDB to learn how to use them.
The MongoDB documents have a pretty simple structure:
{athlete: "Michael Jordan", brand: "Nike"}
{athlete: "Shaquille O'Neal", brand: "Reebok"}
etc..
I want to publish the number of athletes associated with a given brand to a D3.js method to create a piechart. This is easy to do.
The trouble I am having is with collecting brands below a certain threshold (in the example below, brands with only one athlete) into a single row to pass to the client. For example, if the brand distribution is:
{_id: "Nike", count:4} {_id: "Reebok",count:5} {_id: "Puma",count:3}{_id:"Adidas",count:1} {_id:"New Balance", count:1} {_id:"Under Armour", count:1}
I want the output to be:
{_id: "Nike", count:4} {_id: "Reebok",count:5} {_id: "Puma",count:3}{_id:"Others",count:3}
Here is the code I have tried:
pipeline = [
{$group:{
_id: "$brand",
count: {$sum: 1}
}},
{$project:{
_id:{
$cond:[ { $eq: [ "$count", 1 ] }, 'Others', "$_id" ]
},
count:{
$cond: [ { $eq: [ "$count", 1 ] },{$sum:"$count"}, "$count"]
}
}}
];
The _id condition does collapse the single entry brands into 'Others'. If the _id attribute is the only thing passed, I can see the expected results in the browser console (List of brands with more than 1 athletes and Others).
To get the count, I have tried three things:
Setting the true condition in the count if to {$sum:"$count"} as
above
Setting the true condition in the count if to {$sum:1}
Setting count to a non-conditional {$sum:"$count"}
All three result in an empty array being passed to the client. Any ideas on how to get this to work?
Solution:
Setting the pipeline as follows does what I described as desired behaviour in the previous section.
pipeline = [
{$group:{
_id: "$brand",
count: {$sum: 1}
}},
{$group:{
_id:{
$cond:[ { $eq: [ "$count", 1 ] }, 'Others', "$_id"]
},
count:{$sum:"$count"}
}}
];