Informatics Lab Pangeo Workshop 2019

Using Pangeo to solve ‘blockers to the workflow’ in the Met Office.

This post will build upon our previous work on ‘Pangeo’. If you’d like to find out more about the project before reading, please see our previous blog posts.

The Background

We recently had the pleasure of hosting the first Informatics Lab Pangeo Workshop for various Met Office teams spanning many different disciplines.

Blockers to scientific workflows are becomingly increasingly evident in many large-scale organisations, and the Met Office is no exception. More and more, we are engaging in conversations with Met Office employees around big data workflows, covering everything from data discovery, cataloguing and analysis, to publishing and collaboration between teams. The need for open, reproducible and scalable science is becoming increasingly evident by the day.

As with so many large organisations, there are lapses in communication that hinder cross-discipline collaboration. The result: repeatedly we see team after team creating solutions to the same problems, with very little opportunity to share ideas and to learn from each other.

The aims of the workshop were therefore two-fold. Initially, we felt there would be some benefit in the introduction of Pangeo to a wider audience, hoping that some of the concepts would apply to current problems. Secondly, we hoped to enable cross-discipline communication and create an event that would facilitate collaboration between different experts towards streamlining workflows for all, without specialising early on in the process.

The sessions

The outline of the day was carefully curated to encourage creativity and design thinking. The organisation of the sessions were as follows.

1. Introduction to Pangeo (Conceptual as opposed to technical)
2. Defining a use case
3. Identifying workflow blockers
4. A technical tour of Pangeo
5. Hack, collaborate, ideate

Sessions were arranged in this way to encourage design without technical constraint. The day was kicked off with a highly conceptual introduction to Pangeo to set the scene, followed by two design sessions to nail down the current workflow blockers for the identified users, and finally moving on to the introduction of technology that could be applied to designs.

Exercises we felt had the most benefit are listed below:

1. User Profiling: Deciding who you’re targeting in your designs, along with their needs, skills, and biggest frustrations.
2. Task Mapping: Sketching out the journey that leads to meeting acceptance criteria of a particular task.
3. Specifications: Listing the functionality you’d like to have in your end design, from the ‘must-haves’, to the ‘nice-to-haves’.

Example of task mapping

Outcomes

As a direct outcome of the day, we were exposed to some extremely interesting use cases. The areas covered ranged from interactive dashboards for analysing data sets, to tools for the collaboration, and sharing and provenance of big data.

It was also great to listen in on interesting collaborative discussions on shared problems and how we might solve them.

Overall, we hope that the event led to the beginnings of creative thinking around how we could streamline workflows and apply Pangeo principles to specific use cases.