Yerkes Astrophysics Academy for Young Scientists (YAAYS) Hubble data-mining project 10 April 2007 by Vivian Hoette and Max Mutchler

Pick a Hubble dataset to prepare as a high-level science product

Hubble data mining A student / teacher / scientist collaboration in “real data” exploration Coordination and communication via Moodle Basic FITS image display and manipulation (ds9, HOU software) Pick an object; search online archives for available data Determine which images to include/exclude: instruments, filters, exposure times, pointings, etc. Measure shifts need to register (align) all images Combine and clean data Make color-composite image which is visually interesting and conveys more informaton (PhotoShop with FITS Liberator) Submit value-added data to Hubble archive as a “High Level Science Product” (HLSP): more science-ready Data analysis with SciSoft? Or convert to 16-bit for analysis with HOU software? Report analysis and results: papers, posters, talks, etc.

Searching the Hubble data archive

Astronomers using the Hubble telescope have obtained the sharpest view yet of a glowing loop of gas called the Ring Nebula (M57 or NGC 6720), first cataloged more than 200 years ago by French astronomer Charles Messier. The pictures reveal that the "Ring" is actually a cylinder of gas seen almost end-on. Such elongated shapes are common among other planetary nebulae, because thick disks of gas and dust form a waist around a dying star. This "waist" slows down the expansion of material ejected by the doomed object. The easiest escape route for this cast-off material is above and below the star. This photo reveals dark, elongated clumps of material embedded in the gas at the edge of the nebula; the dying central star is floating in a blue haze of hot gas. Ring Nebula

Abstract for O’Dell Workshop poster The newly-formed Yerkes Astrophysics Academy for Young Scientists (YAAYS) is an NSF-funded collaboration of students, teachers, and scientists at Yerkes Observatory. We have identified archival Hubble Space Telescope (HST) observations of the Ring Nebula (NGC 6720 or M57) for an initial excercise in the emerging field of astronomical data mining and curation. The multi-wavelength images of this object from the Wide Field Planetary Camera 2 (WFPC2) made it one of the best-studied planetary nebula, and it's sheer beauty as the first Hubble Heritage release in 1999 made it an instant Hubble icon. In recognition of this dataset's importance to both the study of planetary nebulae and the legacy of Hubble Space Telescope, we chose to prepare and preserve it for posterity. We have collected all available archival WFPC2 data, including some obtained subsequent to 1999, and are converting it into a fully and expertly prepared scientific dataset, using calibrations, methods, and software not available in the 1990s. Our treatment of this dataset will make it more immediately science-ready (and education-ready) than the standard archival products. Further, our prepared dataset will be ingested into the Hubble archive as a High Level Science Product (HLSP), making it queryable by future Hubble and NVO-type data searches. We present our prepared dataset in honor of Bob O'Dell, who has been a central figure in making the Hubble mission a reality, and in making many groundbreaking observations of nebulae with it (including the Ring Nebula).

Ring Nebula project Initial YAAYS data-mining project: an iconic Hubble Heritage dataset (the first one) worthy of careful “curation” for posterity More archival data available to include now, and better software and methods Resulting “high level” data will be more science-ready, and education-ready too YAAYS collaborate on decisions about processing and presenting data Hubble Heritage FITS data release Workshop poster in honor of Bob O’Dell -- former Yerkes Director, Hubble “founding father”, and longtime Heritage collaborator Present final results at ASP Meeting in Chicago?

Example for one filter (F658N): Inspect the input frames: reject any? With ds9 (or HOU): blink-compare images, measure shifts of misaligned stars

Measuring shifts for image registration Display all input images (single_sci.fits) in ds9 with good scaling Find any/all real stars (needles) that appear in all (or most) frames, amid all the artifacts (haystack): cosmic rays, detector artifacts, etc. Click on a particular star to center it, zoom in close, and align all frames (match frames -> image); blink through all images to make sure they are all aligned In each frame, place your cursor over where you think the center of the star is. Stars may be pixelated (not clear exactly where the center is), but by zooming in, you can move you cursor with fractional-pixel accuracy to estimate the x,y center (not the RA,DEC). Use HOU software to measure shifts more precisely? Centroid? Record the x,y position of the star in all 8 frames Pick another suitable star and repeat this process Pick one image to be the arbitrary origin (0,0) and calculate the shift necessary (delta x,y) to get all the other frames aligned to it Repeat this process for all filters (eventually, not for now) Have several students (or groups) independently make measurements Determine the average shifts you measured for all stars; from all students; from all filters. Why are they all slightly different? What shifts should we actually apply?

Stars we can use to measure shifts Image 1

Stars we can use to measure shifts Image 2

Stars we can use to measure shifts Image 3

Stars we can use to measure shifts Sum

Individual WFPC2 image

Sum of all WFPC2 images

Combined clean WFPC2 image

Release data as a High-Level Science Product (HLSP)

Downloading ds9 and FITS data, and making color-composite images SAOimage ds9: MultiMission Archive at STScI (MAST) Hubble Heritage high level science products (HLSP) Sloan Digital Sky Survey (SDSS): NASA Extragalactic Database (NED): FITS Liberator