Publications by Author: Goodman, Alyssa A

• Rule 1. Love Your Data, and Help Others Love It, Too
• Rule 2. Share Your Data Online, with a Permanent Identifier
• Rule 3. Conduct Science with a Particular Level of Reuse in Mind
• Rule 4. Publish Workflow as Context
• Rule 5. Link Your Data to Your Publications as Often as Possible
• Rule 6. Publish Your Code (Even the Small Bits)
• Rule 7. State How You Want to Get Credit
• Rule 8. Foster and Use Data Repositories
• Rule 9. Reward Colleagues Who Share Their Data Properly
• Rule 10. Be a Booster for Data Science

We present Brut, an algorithm to identify bubbles in infrared images of the Galactic midplane. Brut is based on the Random Forest algorithm, and uses bubbles identified by >35,000 citizen scientists from the Milky Way Project to discover the identifying characteristics of bubbles in images from the Spitzer Space Telescope . We demonstrate that Brut's ability to identify bubbles is comparable to expert astronomers. We use Brut to re-assess the bubbles in the Milky Way Project catalog, and find that 10%-30% of the objects in this catalog are non-bubble interlopers. Relative to these interlopers, high-reliability bubbles are more confined to the mid-plane, and display a stronger excess of young stellar objects along and within bubble rims. Furthermore, Brut is able to discover bubbles missed by previous searches—particularly bubbles near bright sources which have low contrast relative to their surroundings. Brut demonstrates the synergies that exist between citizen scientists, professional scientists, and machine learning techniques. In cases where "untrained" citizens can identify patterns that machines cannot detect without training, machine learning algorithms like Brut can use the output of citizen science projects as input training sets, offering tremendous opportunities to speed the pace of scientific discovery. A hybrid model of machine learning combined with crowdsourced training data from citizen scientists can not only classify large quantities of data, but also address the weakness of each approach if deployed alone.

The physical properties of molecular clouds are often measured using spectral-line observations, which provide the only probes of the clouds' velocity structure. It is hard, though, to assess whether and to what extent intensity features in position-position-velocity (PPV) space correspond to "real" density structures in position-position-position (PPP) space. In this paper, we create synthetic molecular cloud spectral-line maps of simulated molecular clouds, and present a new technique for measuring the reality of individual PPV structures. Using a dendrogram algorithm, we identify hierarchical structures in both PPP and PPV space. Our procedure projects density structures identified in PPP space into corresponding intensity structures in PPV space and then measures the geometric overlap of the projected structures with structures identified from the synthetic observation. The fractional overlap between a PPP and PPV structure quantifies how well the synthetic observation recovers information about the three-dimensional structure. Applying this machinery to a set of synthetic observations of CO isotopes, we measure how well spectral-line measurements recover mass, size, velocity dispersion, and virial parameter for a simulated star-forming region. By disabling various steps of our analysis, we investigate how much opacity, chemistry, and gravity affect measurements of physical properties extracted from PPV cubes. For the simulations used here, which offer a decent, but not perfect, match to the properties of a star-forming region like Perseus, our results suggest that superposition induces a  40% uncertainty in masses, sizes, and velocity dispersions derived from 13 CO ( J = 1-0). As would be expected, superposition and confusion is worst in regions where the filling factor of emitting material is large. The virial parameter is most affected by superposition, such that estimates of the virial parameter derived from PPV and PPP information typically disagree by a factor of  2. This uncertainty makes it particularly difficult to judge whether gravitational or kinetic energy dominate a given region, since the majority of virial parameter measurements fall within a factor of two of the equipartition level α   2.

sets, though, interactive exploratory data visualization can give far more insight than an approach where data processing
and statistical analysis are followed, rather than accompanied, by visualization. This paper attempts to charts a course
toward “linked view” systems, where multiple views of high-dimensional data sets update live as a researcher selects,
highlights, or otherwise manipulates, one of several open views. For example, imagine a researcher looking at a 3D volume
visualization of simulated or observed data, and simultaneously viewing statistical displays of the data set’s properties
(such as an x-y plot of temperature vs. velocity, or a histogram of vorticities). Then, imagine that when the researcher
selects an interesting group of points in any one of these displays, that the same points become a highlighted subset in
all other open displays. Selections can be graphical or algorithmic, and they can be combined, and saved. For tabular
(ASCII) data, this kind of analysis has long been possible, even though it has been under-used in Astronomy. The bigger
issue for Astronomy and several other “high-dimensional” fields is the need systems that allow full integration of images
and data cubes within a linked-view environment. The paper concludes its history and analysis of the present situation
with suggestions that look toward cooperatively-developed open-source modular software as a way to create an evolving,
flexible, high-dimensional, linked-view visualization environment useful in astrophysical research.

Abstract. We review an approach to observation-theory comparisons we call \Taste-Testing."
In this approach, synthetic observations are made of numerical simulations, and then both real
and synthetic observations are \tasted" (compared) using a variety of statistical tests. We rst
lay out arguments for bringing theory to observational space rather than observations to theory
space. Next, we explain that generating synthetic observations is only a step along the way to
the quantitative, statistical, taste tests that oer the most insight. We oer a set of examples
focused on polarimetry, scattering and emission by dust, and spectral-line mapping in starforming
regions. We conclude with a discussion of the connection between statistical tests used
to date and the physics we seek to understand. In particular, we suggest that the \lognormal"
nature of molecular clouds can be created by the interaction of many random processes, as can
the lognormal nature of the IMF, so that the fact that both the \Clump Mass Function" (CMF)
and IMF appear lognormal does not necessarily imply a direct relationship between them.

We apply Support Vector Machines (SVMs)—a machine learning algorithm—to the task of classifying structures in the interstellar medium (ISM). As a case study, we present a position-position-velocity (PPV) data cube of 12 CO J = 3-2 emission toward G16.05-0.57, a supernova remnant that lies behind the M17 molecular cloud. Despite the fact that these two objects partially overlap in PPV space, the two structures can easily be distinguished by eye based on their distinct morphologies. The SVM algorithm is able to infer these morphological distinctions, and associate individual pixels with each object at >90% accuracy. This case study suggests that similar techniques may be applicable to classifying other structures in the ISM—a task that has thus far proven difficult to automate.