Clinical Data Explorer

Clinical Data Explorer is a Domino App for exploring, filtering, and visualizing clinical and general datasets.

Overview

Clinical Data Explorer is a dataset analysis application that runs as a Domino App. It lets users browse, filter, and visualize tabular data through an intuitive web interface, and ask questions about their data in plain English. It retrieves data via Domino APIs from datasets and Netapp Volumes accessible to the user, and supports CSV, Parquet, SAS (.sas7bdat), SAS Transport (.xpt), and CDISC Dataset-JSON formats.

Who it’s for

Clinical Data Explorer is aimed at clinical data scientists, statistical programmers, data managers, and reviewers who work with clinical trial data (with particular focus on CDISC dataset standards), as well as anyone who needs a quick, no-code way to explore tabular data on Domino. It is especially useful for life-sciences teams, but it works equally well for general datasets.

Key features

  • Table View: Browse and filter datasets in an interactive table, with built-in missing-values analysis, summary statistics, and distinct-value counts.

  • Explore: Build visualizations, including histograms, X-Y plots, and category bar charts.

  • Chat: Ask questions about your data in natural language (requires an LLM provider to be configured).

  • Friendly Names: Toggle human-readable labels in place of raw column names (for example, for ADaM variables).

  • Shareable views: Copy a link that preserves your current filters so colleagues see the exact same view.

How to use it

  1. Load a dataset. Choose a file from the dataset picker and click Load.

  2. Filter the data. Use + Add Filter to build column/operator/value filters, or use Expression filters written in SAS WHERE, R dplyr, or Python pandas syntax. Click Copy Link to share a filtered view.

  3. Explore visually. Switch to the Explore tab and pick a column for a histogram, or choose X and Y axes for an X-Y plot. Optionally group by a category.

  4. Ask questions. If chat is configured, use the Chat tab to ask things like "What columns are in this dataset?" or "Show me the distribution of AGE."

Table View of a loaded dataset with missing-values analysis
Figure 1. Table View of a loaded dataset with the missing-values analysis panel.
Histogram of a selected column with a category-level bar chart on the side
Figure 2. Histogram of a selected column with a category-level bar chart on the side.

Installation and setup

To install this extension, follow the guidelines provided in the manual install instructions.

Troubleshooting

  • "No datasets found" - Configure the user permissions to access any datasets or Netapp Volumes in the given project context where the extension was opened from.

  • Chat shows "Not Configured" - Set the LLM environment variables; for cloud providers verify LLM_API_KEY.

  • Filters not persisting in links - Ensure "Deep linking and query parameters" is enabled in the Domino App settings.

Advanced settings

Most installs work with the defaults. Use these settings when you need to support larger datasets, higher concurrency, or a specific LLM provider.

LLM and chat configuration

The Chat tab uses an OpenAI-compatible API. By default, the app assumes OpenAI, but it can also point at other providers by changing the base URL and model.

  • OpenAI: Set LLM_API_KEY and optionally set LLM_MODEL (for example, gpt-4o). If LLM_BASE_URL is not set, the app uses OpenAI’s default API URL.

  • Domino LLM Gateway: For customer accounts with Domino’s installable LLM Gateway module, point LLM_BASE_URL at the gateway’s OpenAI-compatible endpoint, set LLM_MODEL to the gateway model/deployment name, and provide LLM_API_KEY if the gateway requires one. This lets the app use centrally governed models without changing app code.

  • Azure OpenAI or other hosted gateways: Set LLM_BASE_URL to the provider’s OpenAI-compatible endpoint, then set LLM_API_KEY and LLM_MODEL as required by that provider.

  • Local Ollama: Set LLM_BASE_URL to an OpenAI-compatible local endpoint such as http://localhost:11434/v1 and set LLM_MODEL. An API key is not required for local Ollama.

  • Chat history: Use CHAT_AGENT_MESSAGE_HISTORY_CACHE_SIZE_B to control the in-memory chat history cache size, and CHAT_AGENT_MESSAGE_HISTORY_CAP to cap how many messages are retained per chat session.

Dataset size, caching, and concurrency

To configure your Domino project settings with regards to dataset size, concurrency, and how downloaded files, MCP server DataFrames, and session metadata are cached, you can follow the guidelines in the Clinical Data Explorer README.md.

Note

The README points to the main branch on GitHub. If you’re on a different release / version of the app, the instructions on main may not match your version. Switch to the corresponding version tag on GitHub that matches your release from the branch / tag dropdown.

Production server settings

See the Clinical Data Explorer README.md for information on configuring your production server settings.