NETWORK.SIMILARITY OpenBI

Finds the nodes that are alike because they have the same neighbours, such as beers bought by the same customers.

Group: Networks. OpenBI’s own: DAX has no function of this name.

Syntax

NETWORK.SIMILARITY ( <links>, <from>, <to> [, <method>] [, <top>] [, <threshold>] )

Parameter	Type	Description
`links`	Table	The rows that make the network, one link a row: a table of the model or a table expression, as the filters leave it.
`from`	Expression evaluated for each row	One end of each row's link, such as a payer, a customer or a place.
`to`	Expression evaluated for each row	The other end of each row's link.
`method`	Text optional	"jaccard" (when left out), "dice", "cosine" or "adamic-adar".
`top`	Whole number optional	How many of each node's most similar nodes are kept: 10 when left out, 0 for every pair.
`threshold`	Number optional	Keep only pairs at least this similar, such as 0.5 for half their neighbours shared (Jaccard). Left out, every pair the top keeps.

Returns

Table. A row a pair, with the columns Node, Other, Shared (the neighbours they have in common) and Similarity.

Remarks

Two nodes are alike when they have the same neighbours, whether or not they are linked themselves: two products bought by the same customers, two customers who buy the same products.
Jaccard is shared / (all the neighbours of either); Dice is 2 × shared / (the sum of their neighbours); cosine is shared / √(the product of their numbers of neighbours); Adamic-Adar adds 1 / ln(neighbours) for each shared neighbour, so one shared with few others counts more.
Only pairs with at least one neighbour in common appear. Each pair appears twice, once from each side; each node's most similar come first.
The links are taken both ways, and no weight is used.
In a network of two kinds of node, such as customers and products, the result holds pairs of both kinds: filter it to the kind you want.

Compared with DAX

An OpenBI function: DAX has none like it, so a measure or query that uses it works in OpenBI only, not in Power BI or Analysis Services.

Examples

Beers bought by the same customers

Pils and IPA share two of their three customers, and so do Stout and Cider.

EVALUATE
VAR Baskets =
    DATATABLE (
        "Customer", STRING, "Product", STRING,
        {
            { "Ola", "Pils" }, { "Ola", "IPA" }, { "Kari", "Pils" }, { "Kari", "IPA" }, { "Kari", "Stout" },
            { "Per", "Stout" }, { "Per", "Cider" }, { "Liv", "Cider" }, { "Liv", "Stout" }, { "Nils", "Pils" }
        }
    )
RETURN
    FILTER (
        NETWORK.SIMILARITY ( Baskets, [Customer], [Product] ),
        [Node] IN { "Pils", "IPA", "Stout", "Cider" }
    )
ORDER BY [Node], [Similarity] DESC, [Other]

Node	Other	Shared	Similarity
Cider	Stout	2	0.666666666666667
IPA	Pils	2	0.666666666666667
IPA	Stout	1	0.25
Pils	IPA	2	0.666666666666667
Pils	Stout	1	0.2
Stout	Cider	2	0.666666666666667
Stout	IPA	1	0.25
Stout	Pils	1	0.2

Each friend's closest match

Adamic-Adar, the best one for each: a friend shared with few others weighs more.

EVALUATE
VAR Links =
    DATATABLE (
        "From", STRING, "To", STRING,
        {
            { "Ann", "Bo" }, { "Ann", "Cy" }, { "Bo", "Cy" }, { "Cy", "Di" },
            { "Di", "Ed" }, { "Ed", "Fay" }, { "Di", "Fay" }, { "Fay", "Gus" }
        }
    )
RETURN
    NETWORK.SIMILARITY ( Links, [From], [To], "adamic-adar", 1 )

Node	Other	Shared	Similarity
Ann	Cy	1	1.44269504088896
Bo	Cy	1	1.44269504088896
Cy	Ann	1	1.44269504088896
Di	Fay	1	1.44269504088896
Ed	Cy	1	0.910239226626837
Fay	Di	1	1.44269504088896
Gus	Di	1	0.910239226626837

Only the close pairs

Every pair, but only those sharing at least two thirds of their neighbours (new in 0.3.77).

EVALUATE
VAR Baskets =
    DATATABLE (
        "Customer", STRING, "Product", STRING,
        {
            { "Ola", "Pils" }, { "Ola", "IPA" }, { "Kari", "Pils" }, { "Kari", "IPA" }, { "Kari", "Stout" },
            { "Per", "Stout" }, { "Per", "Cider" }, { "Liv", "Cider" }, { "Liv", "Stout" }, { "Nils", "Pils" }
        }
    )
RETURN
    NETWORK.SIMILARITY ( Baskets, [Customer], [Product], "jaccard", 0, 0.66 )

Node	Other	Shared	Similarity
Cider	Stout	2	0.666666666666667
IPA	Pils	2	0.666666666666667
Kari	Ola	2	0.666666666666667
Liv	Per	2	1
Ola	Kari	2	0.666666666666667
Per	Liv	2	1
Pils	IPA	2	0.666666666666667
Stout	Cider	2	0.666666666666667

NETWORK.NODES, NETWORK.COMMUNITIES, NETWORK.PREDICT, FILTER.

NETWORK.SIMILARITY is OpenBI's own code: it calls no function of Open Graph, OpenBI's graph library. See the OAX network functions and Open Graph.

Last updated 6 November 2026 · OpenBI Studio 0.3.78