Eyes on the future: optimizing science output for next generation surveys with joint crowdsourced and automated classification techniques

利用公民科学从成像数据中成功地进行“众包”星系分类，清楚地表明有大量的志愿劳动资源。未来的调查还将产生光谱，将光分解成不同的波长，以揭示星系的物理状态。目前，雄心勃勃的项目将扩大志愿者对这些光谱数据的参与，为参与者提供更丰富的体验，并解决其他任何方式都无法完成的星系形成研究。如果处理大量光谱数据的可扩展性问题也可以通过这种方式解决，那么对社区和具有科学素养的公众的影响将是巨大的。虽然众包星系分类已经用斯隆数字巡天（SDSS）十年来的图像证明了它们的价值，但下一代的一些巡天将是光谱的，这大大增加了数据集的复杂性，同时提供了更大的发现空间。虽然拥有数千万个物体的光谱将允许广泛的科学研究，但需要新的工具来最大化科学回报。单独的自动算法要么难以识别困难的特征，要么产生高度污染的样本，试图最大限度地提高完整性。来自即将到来的调查的大量数据使得人工审查变得不切实际。该项目将应对这些挑战，将成功的众包方法扩展到光谱数据，该方法被用于星系动物园项目套件的成像，并扩展了该团队以前由nsf支持的工作。该计划将由两个部分组成：(1)星系托儿所将通过众包光谱数据分类建立一个发射线目录；(2) cluster Scout将识别SDSS中的团块星系。对这些星表的研究将通过以下方式约束星系形成模型：(1)量化具有巨大恒星形成区域的星系的比例；(2)表征这些区域物理性质的内部变化；(3)比较有和没有这些区域的星系的气体金属丰度。与以前的图像工作一样，将要制作的高级目录和将要使用的新分类算法将公开发布，这将是社区的宝贵资源。该研究还将继续在本科天文学课程中实施星系动物园的工作，包括研究生和本科生。

The successful use of citizen science to carry out "crowd-sourced" galaxy classifications from imaging data has shown clearly the availability of a vast resource of volunteer labor. Future surveys will also produce spectra, breaking down light into its wavelengths to reveal the physical state of the galaxies. The present, ambitious, project will extend the involvement of volunteers into these spectroscopic data, providing a richer experience for participants and addressing research in galaxy formation impossible to accomplish any other way. If the scalability problem of handling massive amounts of spectral data can also be solved this way, the impact for the community and for the scientifically literate public will be enormous.While crowdsourced galaxy classifications have proven their worth with a decade of images from the Sloan Digital Sky Survey (SDSS), several next generation surveys will be spectroscopic, which substantially increases the complexity of the data sets, while providing a much larger space for discovery. Although having spectra for tens of millions of objects will allow a wide range of science, new tools are needed to maximize the scientific return. Automatic algorithms alone will either struggle in identifying difficult features or produce highly contaminated samples to try to maximize completeness. The volume of data from forthcoming surveys renders human scrutiny impractical. This project will meet these challenges, extending to spectroscopic data the successful crowdsourcing approach used for imaging by the suite of Galaxy Zoo projects, and expanding previous NSF-supported work by this team. There will be two components: (1) Galaxy Nurseries will build an emission line catalog through crowdsourced classifications of spectroscopic data; and (2) Clump Scout will identify clumpy galaxies in the SDSS. Investigations with these catalogs will constrain galaxy formation models by (1) quantifying the fraction of galaxies with giant star-forming regions; (2) characterizing the internal variation of physical properties of such regions; and (3) comparing the gas metallicity of galaxies with and without these regions.As with the previous image work, the high-level catalogs to be produced, and the new classification algorithms to be used, are to be released publicly, and will be a valuable resource for the community. The study also continues the work to implement Galaxy Zoo in undergraduate astronomy courses, involving both graduate and undergraduate students.