Exploring High-Dimensional Data in Astronomy, Genomics, and beyond, using glue
Publication information:
Abstract
Through an online ViZTIG seminar (Alan Turing Institute), Alyssa Goodman with collaborator Jonathan Foster explains the origins of glue, paying homage to its ancestors DataDesk and Igor, and to cousins like Tableau. Specifically, they show how the high-dimensional nature of data in astronomy, genomics, medicine, and other areas essentially requires a glue-like approach to make the most of the information at-hand.
The great statistician John Tukey highlighted the value of “exploratory data analysis” half a century ago. Since then, not only has the size of data sets researchers might want to explore grown, so has their variety. The analysis/visualization environment known as “glue” is explicitly designed to connect and explore the richness and diversity of modern data, and it is now in use across many fields of inquiry.
The name “glue” is not an acronym. It has three meanings: “glue-ing" together data sets, glue-ing" together charts/plots/tables/maps, and glue-ing" together software tools. The wide variety of “glue” products, all of which are open-source and use python tools, is explained at glue’s and glue solutions' (the glue spinoff) websites, which include video demonstrations.
The talk concludes with a preview of “glupyter,” the browser-based version of glue being developed as a collaboration between glue and Jupyter organizations.
Full text
Through an online ViZTIG seminar (Alan Turing Institute), Alyssa Goodman with collaborator Jonathan Foster explains the origins of glue, paying homage to its ancestors DataDesk and Igor, and to cousins like Tableau. Specifically, they show how the high-dimensional nature of data in astronomy, genomics, medicine, and other areas essentially requires a glue-like approach to make the most of the information at-hand.
The great statistician John Tukey highlighted the value of “exploratory data analysis” half a century ago. Since then, not only has the size of data sets researchers might want to explore grown, so has their variety. The analysis/visualization environment known as “glue” is explicitly designed to connect and explore the richness and diversity of modern data, and it is now in use across many fields of inquiry.
The name “glue” is not an acronym. It has three meanings: “glue-ing" together data sets, glue-ing" together charts/plots/tables/maps, and glue-ing" together software tools. The wide variety of “glue” products, all of which are open-source and use python tools, is explained at glue’s and glue solutions' (the glue spinoff) websites, which include video demonstrations.
The talk concludes with a preview of “glupyter,” the browser-based version of glue being developed as a collaboration between glue and Jupyter organizations.