TOPCAT Visualisation over the Web

TL;DR

This paper introduces a web-based extension of TOPCAT and STILTS, enabling remote interactive visualization of large datasets through a client-server model with efficient pixel-based data transfer.

Contribution

It extends TOPCAT and STILTS to a web client-server architecture, allowing remote, interactive visualization of large tables with minimal bandwidth usage.

Findings

Enables interactive visualization of large datasets over the web.

Supports various plot types and interactions like pan, zoom, and point identification.

Facilitates integration with web and Jupyter environments.

Abstract

The desktop GUI catalogue analysis tool TOPCAT, and its command-line counterpart STILTS, offer among other capabilities visual exploration of locally stored tables containing millions of rows or more. They offer many variations on the theme of scatter plots, density maps and histograms, which can be navigated interactively. These capabilities have now been extended to a client-server model, so that a plot server can be run close to the data storage, and remote lightweight HTML/JavaScript clients can configure and interact with plots based on that data. The interaction can include pan/zoom/rotate navigation, identifying individual points, and potentially subset selection. Since only the pixels and not the row data are transmitted to the client, this enables flexible remote visual exploration of large tables at relatively low bandwidth. The web client can request any of the plot options available from TOPCAT/STILTS. Possible applications include web-based visualisations of static datasets too large to transmit, visual previews of archive search results, service-configured arrays of plots for complex datasets, and embedding visualisations of local or remote tables into Jupyter notebooks.