# Cluster Insights > Cluster Insights - Learn how the Cluster Insights visualization helps you to understand the natural > groupings in your data. This Markdown file sits beside the HTML page at the same path (with a `.md` suffix). It summarizes the topic and lists links for tools and LLM context. Companion generated at `2026-04-24T16:03:56.591480+00:00` (UTC). ## Primary page - [Cluster Insights](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html): Full documentation for this topic (HTML). ## Sections on this page - [View features](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#view-features): In-page section heading. - [Name clusters](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#name-clusters): In-page section heading. - [Add or remove clusters from the display](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#add-or-remove-clusters-from-the-display): In-page section heading. - [Download cluster insights](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#download): In-page section heading. - [Investigate cluster features](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#investigate-cluster-features): In-page section heading. - [Numeric features](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#numeric-features): In-page section heading. - [Categorical features](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#categorical-features): In-page section heading. - [Text features](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#text-features): In-page section heading. - [Image features](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#image-features): In-page section heading. - [Geospatial location features](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#geospatial-location-features): In-page section heading. ## Related documentation - [Classic UI documentation](https://docs.datarobot.com/en/docs/classic-ui/index.html): Linked from this page. - [Modeling](https://docs.datarobot.com/en/docs/classic-ui/modeling/index.html): Linked from this page. - [Model insights](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/index.html): Linked from this page. - [Understand](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/index.html): Linked from this page. - [Feature Impact](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/feature-impact-classic.html): Linked from this page. - [clustering model](https://docs.datarobot.com/en/docs/classic-ui/modeling/special-workflows/unsupervised/clustering.html#build-a-clustering-model): Linked from this page. - [n-grams](https://docs.datarobot.com/en/docs/reference/glossary/index.html#n-gram): Linked from this page. ## Documentation content # Cluster Insights With the Cluster Insights visualization, you can understand and name each cluster in a dataset. Use clustering to capture a latent feature in your data, to surface and communicate actionable insights quickly, or to identify segments in the data for further modeling. > [!NOTE] Note > The maximum number of features computed for Cluster Insights is 100. The features are selected from the features used to train the model, based on the [Feature Impact](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/feature-impact-classic.html) (high to low). The remaining features (those not used to train the model) are sorted alphabetically. To analyze the clusters in your data: 1. Build aclustering modeland expand the model you want to investigate. 2. SelectUnderstand > Cluster Insights. The following table describes the Cluster Insights visualization. ElementDescription1Select clustersClick toselect clustersto view or remove from view.2Rename clustersName clustersafter you gain an understanding of what they represent.3Feature ListBy default, DataRobot builds clustering models using the Informative Features list, although you can select another feature list to compare other features. Analyzing features not used to generate the clusters can still be useful, for example, to answer questions like "How does income distribute among my clusters, even if I'm not using it for clustering?"4Download CSVClick to download the cluster insights. The CSV contains the information displayed in the Cluster Insights visualization, and more detailed feature data.5Feature page controlPage through toview more features.6ClustersClusters display in columns of features (four features display by default). Cluster sizes are shown above (in percentages). Clusters are sorted by size from largest to smallest.7Cluster arrowClick to view more clusters. The rightmost cluster contains 100% as a baseline comparison.8FeaturesFeatures are sorted by feature importance. The Informative Feature list displays by default, but you can select another feature list. 3. Evaluate the distribution of descriptive features across clusters andthe feature values in each cluster. ## View features The features display in order of Feature Impact (most important to least). To page through the features, click the right arrow above the clusters: The display defaults to four features but you can view 10 features at a time by clicking the feature page control and selecting 10: ## Name clusters Once you get a sense of what your clusters represent, you can name them. Take a look at the data for obvious similarities and then name the cluster accordingly. The cluster names propagate to other insights and predictions, allowing you to further analyze the clusters. 1. ClickRename clustersand enter names for each cluster. 2. ClickFinish editingand clickProceed. ## Add or remove clusters from the display 1. ClickSelect clustersto choose clusters to view or delete. 2. Click the down arrow to select a new cluster. 3. Click+ Add clusterto display additional clusters. 4. Click the trash can icon to remove a cluster from the display. ## Download cluster insights You can download the cluster insights as a CSV file for further analysis by clicking Download CSV above the clusters. ## Investigate cluster features The following sections show the visualization tools used to investigate cluster features. The sample dataset contains features representing housing data: This dataset could be run in supervised mode with `price` as the target feature, but for clustering mode, no target is specified. The dataset contains the following feature types: - Numeric ( price , sq_ft , etc.) - Categorical ( cooling , roof_type , etc.) - Text ( amenities ) - Image ( exterior_image ) - Geospatial ( zip_geometry ) ### Numeric features To view numeric features in Cluster Insights: 1. Locate a numeric feature on theCluster Insightstab. 2. Click near the feature name to expand. Hover over the blue bar for each cluster to view the maximum, median, average, minimum, the percentage missing, and the 1st and 3rd quartiles. ### Categorical features To view categorical features in Cluster Insights: 1. Locate a categorical feature on theCluster Insightstab. Low-frequency labels for the feature are grouped in the Other category. For example, if only a small number of houses in the dataset havefloor_typeengineered wood, houses with engineered wood would be grouped into the Other category for thefloor_typefeature. 2. Hover over the bar for each cluster to see the breakdown within a cluster. 3. Click near the feature name to expand. This allows you to see more categories. 4. To drill into the categories, click the gear icon next to the feature name and selectHigh cardinality view. Hover to see the percentage of records that have each value. ### Text features For text features, Cluster Insights shows [n-grams](https://docs.datarobot.com/en/docs/reference/glossary/index.html#n-gram) ranked by importance (highest to lowest). These are displayed as blue bars that represent the relative importance. To see the actual importance value, [download the CSV](https://docs.datarobot.com/en/docs/classic-ui/modeling/analyze-models/understand/cluster-insights-classic.html#download). 1. Locate a text feature on theCluster Insightstab. Click near the feature name to expand. NoteMissing values impute blanks;blankis included as an n-gram if the missing values are scored as important. 2. Hover over an n-gram in a cluster to view sample strings that contain the word. 3. ClickSee more context examplesto drill down. The Context window displays ten random excerpts that contain the n-gram. ### Image features For image features, Cluster Insights displays sample images from each cluster. DataRobot uses the [Maximal Marginal Relevance](https://www.cs.cmu.edu/~jgc/publication/The_Use_MMR_Diversity_Based_LTMIR_1998.pdf) criterion to choose images that are representative of the cluster, but also diverse within the cluster (so not all from the [centroid](https://docs.datarobot.com/en/docs/reference/glossary/index.html#centroid) of the cluster). 1. Locate an image feature on theCluster Insightstab. By default, four images are displayed. Click near the feature name to show 10 images. 2. Hover over an image to zoom in. ### Geospatial location features To see a map of a geospatial location feature: 1. Locate a geospatial feature on theCluster Insightstab. DataRobot uses theMaximal Marginal Relevancecriterion to transform geospatial data to points. TipDataRobot derives numeric features (such as area and coordinates) from geospatial features. Often the derived features appear in theInformative Featureslist, while the original geospatial feature does not. To view the geospatial map of the original geospatial feature, selectAll Featuresfrom the Feature List dropdown and locate the feature. 2. Click near the feature name to expand the map: To view individual clusters, click theMap legendand click cluster names to hide clusters. The map visualization includes zoom buttons.