Skip to main content
SimilarExport ZIP
100%

A worked example of using the undocumented DAX function KMeansClustering

✓ Published0🌍 Public
HHugoberry
Last edited Jan 18, 2024
Created on Jan 18, 2024

This example demonstrates how to cluster a set of labeled numeric values using the undocumented DAX function `KMeansClustering`, which partitions the sample data into groups based on their attribute values. The visualization shows the resulting cluster assignments for each item, with the function called directly in a DAX query. The code supplies a table of letter-value pairs, specifies the column mappings for the attribute and item labels, and requests cluster IDs as output. The clustering runs with a parameter of -1, indicating automatic determination of the number of clusters, and the result appears as a table of items paired with their assigned cluster identifiers.

AI-generated description

Similar vizzes

Loading thumbnail…

K-means clustering in DAX

This example demonstrates a K-means clustering implementation written in DAX, showing how to group items into clusters based on multiple numeric attributes. The code computes a transformed table using `SUMMARIZECOLUMNS` to aggregate attribute values per item, then passes that table to a custom `KMeansClustering` function. The function assigns each item a `ClusterId` while preserving the original item labels. The visualization highlights the resulting clusters, with the DAX code serving as the core analytical engine rather than relying on a JavaScript library. The source is a gist by Hugoberry, and the rendering approach is not visible here, so the focus remains on the clustering logic and the data transformation pipeline.

HHugoberry
85% match
Loading thumbnail…

KMeans clustering call in DAX Power BI desktop 3D

This example demonstrates a 3D KMeans clustering visualization built directly in DAX for Power BI, using the Iris dataset’s numeric columns (Column6, Column7, Column8). It shows how raw table columns are summarized and then passed into the custom `KMeansClustering` DAX function, which assigns each data point to a cluster and returns a cluster ID alongside the original attributes. The code explicitly constructs a DAX query with `SUMMARIZECOLUMNS` and `SELECTCOLUMNS` to shape the data, then calls `KMeansClustering` with parameter lists defining attribute roles and output columns. The rendering approach is not visible here, but the focus is on the algorithmic clustering step within Power BI’s DAX engine.

HHugoberry
77% match
Loading thumbnail…

Gist 599b117350a329c9046f68ae0ce1e21a

This example visualizes a k-means clustering result over a set of data points, showing how the algorithm partitions the space into distinct clusters based on two measured attributes. The chart is generated from a Power BI DAX query that first summarizes the source table `Query1` by summing the X and Y values for each category Z, then feeds those aggregated coordinates into the `KMeansClustering` function. The output assigns a `ClusterId` to each point, which the visualization likely encodes through color or position. The rendering relies on Power BI’s built-in visual engine rather than a custom JavaScript library, as no D3 or Canvas code appears in the provided DAX.

HHugoberry
75% match
Loading thumbnail…

R flavoured Kmeans algorithm in Power Query

This example demonstrates the R-flavored K-means clustering algorithm executed directly within Power Query, using the iris dataset’s petal width and length columns to assign cluster memberships. The visualization shows the resulting iris data points colored by their assigned cluster, with the clustering computed in R via the `kmeans` function and seeded for reproducibility. The Power Query M code invokes the R engine through the `R.Execute` API, passing a script that runs K-means with three centers and twenty random starts, then returns the augmented table containing the cluster labels. The underlying data originates from the built-in iris dataset, and the rendering relies on Power Query’s native table output, not a separate charting library.

HHugoberry
63% match
Loading thumbnail…

K means clustering

This example shows the k-means clustering algorithm in action, animating the iterative process of grouping a small set of points from `data.csv` into three clusters. Initially, three randomly placed centroids are drawn, and each iteration assigns points to the nearest centroid using Euclidean distance, then moves the centroids to the cluster mean. The visualization uses D3 v6’s selection and `transition` methods to animate the points and centroids, with the centroid styled distinctly via a black stroke. The iteration count is displayed as text, and the animation stops after ten iterations.

Kkaviya-ds
62% match
Loading thumbnail…

KMeans clustering

This example demonstrates KMeans clustering on a small dataset of six points loaded from a CSV file. It shows how points are iteratively assigned to one of three clusters, with colored hulls drawn around each group and centroids recalculated until a local minimum cost is reached. The visualization uses D3.js v4 with a React framework, rendering SVG circles and polygon hulls with animated transitions. The code employs `d3.csv`, `d3.polygonHull`, `d3.scaleLinear`, and `d3.interval` to animate the clustering process, displaying the iteration count in the final state.

NNaveenaS12
60% match
Loading thumbnail…

Clusters

This interactive scatterplot visualizes galaxy data with K-Means clustering, allowing users to switch between original magnitude coordinates, principal component projections, and t-SNE embeddings. The visualization supports dynamic axis selection from five photometric bands, with points colored by cluster membership across different projection methods. Built with d3.v3, the chart uses SVG circles, d3.scale.linear for axes, and d3.svg.axis to render coordinates, with clickable buttons and dropdowns to update the display. Data comes from finalOutput.csv containing astronomical measurements and cluster assignments.

JJverma
60% match
Loading thumbnail…

K-means clustering of wine data

This visualization demonstrates the results of K-means clustering on the wine dataset, plotting each sample as a dot based on its Alcohol and Malic attributes, with colors indicating cluster membership. The code uses D3.js v3 to render an SVG scatterplot, employing d3.scale.linear for axes, d3.svg.axis for tick marks, and d3.scale.category20 for coloring clusters. Mouseover events trigger a tooltip showing cluster, Alcohol, and Malic values. Data comes from the clusters.csv file, which includes precomputed cluster assignments.

JJverma
57% match