Breakout Summary Report

ARM/ASR User and PI Meeting

19 - 23 March 2018

Session Title: Tutorial - Interacting with ARM radar data using Python
Session Date: 19 March 2018
Session Time: 9:00 AM - 11:30 AM
Number of Attendees: 18
Summary Authors: Scott Collis

Breakout Description

Python ARM Radar Toolkit short course at the ASR/ARM Science Team Meeting

Main Discussion

The Python ARM Radar Toolkit (Py-ART) [1] is a Python-based architecture for working with weather radar data. Built to facilitate rapid implementation of radar science into ARM infrastructure and benefiting from ARM support, Py-ART has hundreds of users worldwide and over 19,000 downloads.

Given the critical mass of stakeholders, the development team decided to hold a short course on the morning of the meeting. Attendance was by registration and the registration processed captured the radar and python proficiencies of attendees. Attendees were generally weak-to-new Python users and none had used the Py-ART before. To this end the instructors (Scott Collis and Robert Jackson [ANL] and Valentin Louf [Monash University]) designed the course to be 70/30 introduction to Python/Py-ART.

Key Findings

The course had 17 sign-ups and 18 attendees. Attendees were a broad mix of University PIs, students, postdocs, and ARM laboratory staff (including a healthy contingent from the ARM Data Center). Coursework consisted of working through Jupyter notebooks [2] introducing attendees to Python, Scientific Python, and then Py-ART. For the Py-ART component of the course users interacted with X-band Scanning ARM Precipitation Radar (XSAPR) data from the Southern Great Plains (SGP) site. Attendees learned how to read XSAPR data, plot data on a map of the SGP, and do field-based calculations and save to a community format-compliant file. In addition, users were shown how they could combine ARM radar data with radar data from NOAA by access directly from Amazon Web Services. In the future we will be setting up a similar demonstration using the ADC’s new RESTful download interface.

Decisions

Despite several hiccups and the very short nature of the course we received very positive feedback from attendees and have been in contact with several post course. As part of the course we signed attendees up to the Py-ART users Slack workspace [3] thus enabling future engagement.

Action Items

[1] https://github.com/ARM-DOE/pyart
[2] https://github.com/EVS-ATMOS/stm_2018_pyart_course
[3] https://pyart-users.slack.com/ (sign up needed. Contact Scott Collis if interested)