Run Principal Component Analysis in Your Browser - instaPCA

instaPCA runs principal component analysis (PCA) entirely in your browser. CSV data is read and processed locally with JavaScript, so it is never uploaded.
The default standardized analysis is comparable to prcomp(scale.=TRUE) in R. Additional options let you adjust centering, scaling, and covariance behavior.

1. CSV Loading

Load CSV Not selected

Transpose (row/column swap):

Variable names row:
Sample names column:

* Preview shows maximum 10 rows × 10 columns

Variables: -
Samples: -
Principal Components: -

Analysis Method:

Covariance (with standardization) Covariance (centering only) No centering/standardization

Details (click to expand)

Covariance (with standardization): Standardizes each variable to mean 0 and variance 1 before PCA. This is equivalent to correlation-matrix PCA, such as R's prcomp(..., scale.=TRUE) or Python PCA after StandardScaler. It is useful when variables use different units.
Covariance (centering only): Centers each variable at mean 0 and runs PCA on the covariance matrix. This matches the default behavior of R's prcomp and Python's sklearn.decomposition.PCA. Variables with larger variance will have more influence.
No centering/standardization: Uses the original values without transformation, corresponding to R's prcomp(center=FALSE, scale.=FALSE). This is generally not recommended, but can be appropriate for data that has already been centered.

Re-standardize scores after PCA (optional)

Use population covariance (denominator n instead of n-1)

Details (click to expand)

Re-standardize scores after PCA (optional): Runs PCA after centering, then rescales the principal component scores to variance 1.
Use population covariance (denominator n instead of n-1): Uses the sample count n as the covariance denominator instead of the usual sample-covariance denominator n-1. Leave this off for results comparable to the standard defaults in R and Python. Enable it only when the data represents the complete population.

Preview displays up to 1000x1000. Export includes all data except plot data.

instaPCA is intended for exploratory analysis and learning. Do not use it as the sole basis for medical decisions or other high-impact business decisions.
Check that the content and structure of your input data are appropriate before relying on the results.

The site operator and tool creator are not liable for losses or damages arising from use of this tool.
Verify the output carefully. You are responsible for decisions and actions based on the results.
The tool and its availability may change without notice.