This specification describes an extension for MapConfig 1.4.0 version.
This extension depends on Analyses extension. It extends MapConfig with a new attribute: dataviews
.
It makes possible to get tabular data from analysis nodes: aggregated lists, aggregations, and histograms.
An aggregation is a list with aggregated results by a column and a given aggregation function.
Definition
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
{
// REQUIRED
// string, `type` the aggregation type
“type”: “aggregation”,
// REQUIRED
// object, `options` dataview params
“options”: {
// REQUIRED
// string, `column` column name to aggregate by
“column”: “country”,
// REQUIRED
// string, `aggregation` operation to perform
“aggregation”: “count”
// OPTIONAL
// string, `aggregationColumn` column value to aggregate
// This param is required when `aggregation` is different than "count"
“aggregationColumn”: “population”
}
}
Expected output
1
2
3
4
5
6
7
8
9
10
11
12
13
{
"type": "aggregation",
"categories": [
{
"category": "foo",
"value": 100
},
{
"category": "bar",
"value": 200
}
]
}
Histograms represent the data distribution for a column.
Definition
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
{
// REQUIRED
// string, `type` the histogram type
“type”: “histogram”,
// REQUIRED
// object, `options` dataview params
“options”: {
// REQUIRED
// string, `column` column name to aggregate by
“column”: “name”,
// OPTIONAL
// number, `bins` how many buckets the histogram should use
“bins”: 10
}
}
Expected output
1
2
3
4
5
{
"type": "histogram",
"bins": [{"bin": 0, "start": 2, "end": 2, "min": 2, "max": 2, "freq": 1}, null, null, {"bin": 3, "min": 40, "max": 44, "freq": 2}, null],
"width": 10
}
Formulas given a final value representing the whole dataset.
Definition
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
{
// REQUIRED
// string, `type` the formula type
“type”: “formula”,
// REQUIRED
// object, `options` dataview params
“options”: {
// REQUIRED
// string, `column` column name to aggregate by
“column”: “name”,
// REQUIRED
// string, `aggregation` operation to perform
“operation”: “count”
}
}
Operation must be: “min”, “max”, “count”, “avg”, or “sum”.
Result
1
2
3
4
5
6
{
"type": "formula",
"operation": "count",
"result": 1000,
"nulls": 0
}
The new dataviews attribute must be a dictionary of dataviews.
An analysis node id can be referenced from dataviews to consume its output query.
The layer consuming the output must reference it with the following option:
1
2
3
4
5
6
7
8
9
{
// REQUIRED
// object, `source` as in the future we might want to have other source options
"source": {
// REQUIRED
// string, `id` the analysis node identifier
"id": "HEAD"
}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
{
"version": "1.4.0",
"layers": [
{
"type": "cartodb",
"options": {
"source": {
"id": "HEAD"
},
"cartocss": "...",
"cartocss_version": "2.3.0"
}
}
],
"dataviews" {
"basic_histogram": {
"source": {
"id": "HEAD"
},
"type": "histogram",
"options": {
"column": "pop_max"
}
}
},
"analyses": [
{
"id": "HEAD",
"type": "source",
"params": {
"query": "select * from your_table"
}
}
]
}
Camshaft’s analyses expose a filtering capability and aggregation
and histogram
dataviews get them for free with
this extension. Filters are available with the very dataview id, so if you have a “basic_histogram” histogram dataview
you can filter with a range filter with “basic_histogram” name.
Allows to remove results that are not contained within a set of elements.
Initially this filter can be applied to a numeric
or text
columns.
Params
1
2
3
4
{
“accept”: [“Spain”, “Germany”]
“reject”: [“Japan”]
}
Allows to remove results that don’t satisfy numeric min and max values. Filter is applied to a numeric column.
Params
1
2
3
4
{
“min”: 0,
“max”: 1000
}
Filters must be applied at map instantiation time.
With :mapconfig as a valid MapConfig and with :filters (a valid JSON) as:
GET /api/v1/map?config=:mapconfig&filters=:filters
POST /api/v1/map?filters=:filters
with BODY=:mapconfig
If in the future we need to support a bigger filters param and it doesn’t fit in the query string, we might solve it by accepting:
POST /api/v1/map
with BODY={“config”: :mapconfig, “filters”: :filters}
Assume :params (a valid JSON) as named maps params, like in: {“color”: “red”}
GET /api/v1/named/:name/jsonp?config=:params&filters=:filters&callback=cb
POST /api/v1/named/:name?filters=:filters
with BODY=:params
If, again, in the future we need to support a bigger filters param that doesn’t fit in the query string, we might solve it by accepting:
POST /api/v1/named/:name
with BODY={“config”: :params, “filters”: :filters}
A bounding box filter allows to remove results that don’t satisfy a geospatial range.
The bounding box special filter is available per dataview and there is no need to create a bounding box definition as it’s always possible to apply a bbox filter per dataview.
A dataview can get its result filtered by bounding box by sending a bbox param in the query string,
param must be in the form west,south,east,north
.
So applying a bbox filter to a dataview looks like: GET /api/v1/map/:layergroupid/dataview/:dataview_name?bbox=-90,-45,90,45