inst/helpfiles/sidebar_help.md

In rhenkin/visxhclust: A Shiny App for Visual Exploration of Hierarchical Clustering

Sidebar

This sidebar controls the main clustering workflow:

1. Load an RDS, CSV, TSV or tab-delimited text file
2. Select a scaling method, distance metric and linkage method
3. Select the desired number of clusters
4. Select the variables/features to use in clustering

Most tabs will then be updated without any additional inputs. The sidebar contains two other controls to help with the analysis:

Heatmap features: variables selected from this list will be added to the top of the heatmaps annotations. This list will automatically include any non-numeric variable loaded with the data (e.g. text) or numeric variables that contain missing values, which cannot be used for clustering.

Correlation-based feature selection: a one-sided t-test based on the chosen threshold will be use to automatically remove highly correlated variables. You can inspect the relationship between the variables that were deselected in the Data overview tab, under Highly correlated variables.

Example data

The following datasets can be downloaded, inspected and loaded into the tool to understand the required format:

1. Normally-distributed data
2. Normally-distributed data with missing values
3. Normally-distributed data with annotation
4. Binary data
5. Log-scaled data

rhenkin/visxhclust documentation built on March 28, 2023, 8:16 p.m.