Repo for an upcoming paper analyzing gravitationally unbound, coherent structures with significant velocity gradients in nearby star forming molecular clouds. The repo is prepared to work with binder and will be shared alongside the paper on Authorea. By working between Github and other data/code/article sharing services, we hope to present what a journal article could look like in an era of open data and reproducible/reusable codes.
Please contact Hope Chen at hhchen@cfa.harvard.edu if you have any questions. Any suggestions are also welcome via the issue tracker of this repo.
The project is lead by Hope Chen at Harvard-Smithsonian Center for Astrophysics, and is a collaboration with Jaime Pineda, Alyssa Goodman, and Andreas Burkert.
Using data from the GBT Ammonia Survey (GAS) Data Release 1 (Friesen and Pineda et al., 2017), we look into the physical properties of structures that are coherent (previous examples in Goodman et al., 1998 and in Pineda et al, 2010) and show significant velocity gradients (previous examples in Goodman et al., 1993). With a much improved physical resolution of ~4000 AU (compared to ~0.1 pc in the 1990s) at the distance of Ophiuchus and Taurus, one goal of the analysis is to provide updated numbers for properties of (potentially) rotational motions within 0.1 pc, which has been essential in simulations and analytic models alike in the past two decades. In our first analysis, these cores seem to be gravitationally unbound and are termed "droplets." In the paper, we hope to provide a sensible guess to what role these structures could play in the process of star formation.
The main dataset used in the analyses is from the GAS DR1 (Friesen and Pineda et al., 2017) and is hosted on Dataverse. The Herschel column density maps are derived by Ayushi Singh and Peter Martin at University of Toronto, using archival data from the Herschel Space Observatory. The data in this repo are copies of the data hosted on Dataverse, without any modification.
Due to the Github policy, data files larger than 100 MB are not hosted here. These include the raw data cubes and the position-position-velocity cubes based on results of Gaussian line fitting to the ammonia hyperfine lines. These large files are not needed to run codes in this repo. Please look in the GBT DR1 dataverse for the files.
A manuscript is being prepared on Authorea. Meanwhile, results of analyses are actively shared in this repo, together with the code used for these analyses. Currently, the codes are not yet stable enough to be applied on other datasets. Please approach with caution.
It has become fashionable to share data and codes in the fast developing field of data science in the study of natural sciences. However, the lack of a clear guideline for data sharing has often resulted in difficulties to reproduce (and, importantly, test) the results published using the original version of the shared data, and to reuse the shared data in other analyses.
We find an example of a complete set of guidelines presented by Ellis & Leek very useful. Ellis & Leek list the following as "the information you should pass to a statistician":
- The raw data
- A tidy data set
- A code book describing each variable and its values in the tidy data set
- An explicit and exact recipe you used to go from 1 -> 2, 3. (sic)
In this project, we try to follow these guidelines, by: 1) pointing to the raw datasets in the original surveys, 2) storing the tidied datasets in this github repo, and 3) providing the code book and the recipe in the format of ipython notebooks. We hope that, by doing so, this project could be an example of data sharing for future scientific projects.