The Eternal Sunshine
of Open-Science
$ whoami
- Member of Technical Staff @ LambdaTest
- GSoC '23 @ OpenAstronomy/SunPy
- Volunteer @ OSDC, FOSS United
- (Occasional ;_;) OSS Contributor
- Interested in Distributed Systems, Python, Scientific Software
an outline
- Open-Science
- FOSS and Science
- Getting involved
two researchers
-
Peter unable to reproduce results due to
-
Tony uses open datasets, FOSS libraries, and collaborative infra to publish reproducible research globally.
😨



The Sun
best star ever 🌟
Why study the Sun?

many reasons
It powers all life on Earth!
Space Weather
Climate
Physics, duh
...
But the Sun is also a stormy place...

SDO/AIA
In a sky full of stars...


SDO/AIA

SDO/AIA
Coronal Mass Ejections

SolarSoft
Great tool!
Built on IDL, proprietary
Closed development environment
Lack of versioning, relatively silo'ed
The SunPy Project
A community of scientists and software developers creating an ecosystem of Python packages for solar physics.

The SunPy core Package
Solar data analysis. This involves:
- access data from different providers
- read image and time series data
- visualise, plot solar data
- transform between commonly used coordinate systems
- ...
Objective
To develop and facilitate a high-quality, easy-to-use, community-led, FOSS solar data analysis ecosystem based on the scientific Python environment.
SunPy Affiliate Projects
- Sunkit-image: Advanced image processing for solar data
- NDCube: Multi-dimensional data structures for scientific data
- drms: Access to Stanford's Joint Science Operations Center data
- ... many more
SunPy's Impact
- Used in numerous scientific publications
- Supports various solar missions and ground-based observatories
- Enables efficient data analysis and visualization for researchers worldwide
A quick tour!
- Units
- Times
- Coordinates
- Maps
- Timeseries
Units...
- sunpy is always explicit about the units associated with a value
- utilise
astropy.units.Quantity
objects
Times and TimeRanges

Acquiring Solar Data
- Support for available data sources: VSO, HEK, JSOC...
- Fido, SunPy's unified search interface
Acquiring Solar Data



very extensible!
Adding Sources
- Just need to subclass existing BaseClients, map API's

Plotting a Solar Map

Working with Maps
Finding bright regions via ndimage

https://docs.sunpy.org/en/stable/generated/gallery/map
Open-Science
- the movement to make scientific research and its results accessible to all levels of society.
- transparency, reproducibility, collaboration.
- not just open access publishing

COVID 19


“We have seen the problems caused by the lack of data sharing on COVID-19,” said Dr. Soumya Swaminathan, WHO Chief Scientist. “When data related to research activities are shared ethically, equitably and efficiently, there are major gains for science and public health.”
Open-Scie☂️ce
Open Data
Open Peer Review
Open Access
Open Resources
Citizen Science
DEI
Open Software
Open Results
Open Hardware
FAIRness in Data
- Findability
- Accessibility
- Interoperability
- Reproducibility
Research at large!
Open Sci Infra




rethinking research outputs
- Rep is really important in research
- Traditional science = results oriented
More than just results
-
Datasets
-
Methods or Protocols
-
Educational resources
-
Software tools
-
Code
More than just openness
-
Openness ≠ Rigor
-
Need high-quality, trustworthy research
Where FOSS comes in
-
Accessibility
-
Transparency
-
Encourages community contributions
a familiar example

sunpy's impact
- Used in numerous scientific publications
- Supports various solar missions and ground-based observatories
- Enables efficient data analysis and visualization for researchers worldwide
Acquiring Solar Data


Adding Sources
- Just need to subclass existing BaseClients, map API's

very extensible!
🗣️ Another One


JupyterCAD

FOSS
Science
Scientific challenges inspire better FOSS
FOSS enables better science
Scientific Python
Python 🤝 Scientific Communities!
- Accessible
- Flexible
- Open-Source
- Actively Developed
- Strong Community Support
Many Reasons Why








Some blindspots
-
Open Source tools with no docs
-
Datasets without metadata
-
Institutions Open-Sourcing; never accepting PRs
Challenges to Open-Science
-
Project sustainability
-
Lack of incentives
-
Scientific rigor in fast-evolving projects
The AI age
- Model reproducibility crisis
Policy demands for transparency
"Open-Source AI"
Getting Involved
(as a researcher)
-
Publish with reproducibility tools
-
Use open datasets/APIs
-
Join communities





Getting Involved
(as a developer)
-
Contribute to FOSS projects in science
-
University programs for students
-
Opportunities like GSoC, Outreachy, ESA/NumFOCUS projects
Relevant links
- https://github.com/sunpy/
- https://docs.sunpy.org/
- https://book.the-turing-way.org/
- https://reproducibilitea.org/
- https://nasa.github.io/Transform-to-Open-Science/
in/akshit-tyagi42

exitflynn

these slides ^

flynnzler

@exitflynn@fosstodon.org

Copy of All is (not) FAIR
By Akshit Tyagi
Copy of All is (not) FAIR
- 11