5 Minutes to Spectra Sherpa Cloud

0.6 trial

The hosted trial uses a supplied starter project and a qualified registered-reference path. It is intended for evaluation, not for confidential or regulated customer data.

Use this path to run the supplied Lavender starter or an exact reference file you download directly from its provider. The hosted trial does not retrieve or redistribute the provider file.

Use the published trial link supplied by Spectra Scientific. Sign up with your individual email address and complete the single-use six-digit passcode check. For the rate, session, and scientific-input boundaries, see Demo Access and Limits.

2. Run an Example PCA with New Analysis

For Lavender, open New Analysis and use its supplied project. For a third-party reference, first open Data > My Dataset > Reference Datasets, choose Browse all Eigenvector datasets, download Corn in the provider's new tab, then choose Import downloaded file. After the exact size and SHA-256 match succeeds, bind that dataset from Use My Dataset and run the analysis.

Ask Advisor along the way

Advisor can answer questions at any step: "What does this score plot suggest?", "Which preprocessing should I try next?", "Are these samples outliers?". Use it as an interpretation aid, not a replacement for scientific judgment.

A useful first question

Ask Advisor to identify the dataset, target, units and population it is discussing before asking for a recommendation. Junior users can start with “What is this plot showing?” Senior users can ask for competing explanations and the diagnostic that would distinguish them.

3. Open the Detailed Node Viewer

Click into a result node to open the Node Detail view. This is where you read a step end to end: the plot, the table, the metrics, and the diagnostics for that node. Click through the workflow nodes to see how the data changes from import to model output.

4. Inspect the Supplied Dataset

Open Data and inspect the admitted files, metadata, and data matrix. The trial accepts no general customer files; the reference selector accepts only an exact registered file that the user obtained independently. For other scientist-owned data, use Local OSS, Subscription Cloud, or Enterprise Hybrid.

5. Build a PCA Variant

Duplicate the example sheet while retaining its bound starter source. Adjust preprocessing and components, run it, and use Node Detail to review the result. The worker re-admits the exact starter grant before every execution.

6. Export and Report

Export the results, including figures, tables, or the workflow, and open the Report for a shareable record of the run. If an AI narrative is available, treat it as a labeled interpretation draft; the plots and metrics remain the scientific record.

What to Remember

Trial accounts are for evaluation on supplied data. Customer scientific input is refused. The canonical policy page is Demo Access and Limits.