6 ms·

Caught for the First Time: The Early Flash of an Exploding Star

ezequiel-garzon 11y ago
If I understand correctly, this data was recorded in 2011. It amazes me that it could take NASA this long to sift through its own (massive) data. As datasets are getting ever larger (outpacing gains in computing power, I feel) I wonder if some similar discoveries will take decades (centuries?) in the future. Does anybody know if these teams make it a point to keep datasets "small enough" to be manageable in reasonable time frames?
- cyphar 11y ago
  It should be noted that NASA makes essentially all of it's data public. In the astrophysics community, this means that researchers write scripts to run over all of the data for exploratory research and then do more complicated analysis later. It's also important to note that one of the best and most accurate space telescopes ever made (Kepler) was keeping the entire photometric community very busy.
  - ezequiel-garzon 11y ago
    Thanks for your comment! Do you feel that, in the medium term, as ever larger datasets are generated the overall computing power will be unable to catch up?
    - greglindahl 11y ago
      Data gets bigger because instruments have bigger/better sensors (the biggest of which are built like computers) and there is more storage and bandwidth. So it basically advances in lockstep with computing power. There's a lot of hand-waving and details, but that's the overall answer.
    - cyphar 11y ago
      To be quite frank, the limiting factor is not computing power, it's code written by physicists. I'm a software developer by trade, and it's sometimes painful to see the type of code my supervisors write. I get the feeling that "our simulation takes half an hour" is acceptable and they won't try to optimise it unless they're near a deadline. Obviously, that's not deriding them, they have much better things to do than to micro-optimise their code. But in general I prefer writing all of my analysis scripts from scratch after discussing what methods they used. :P
  - bartman 11y ago
    Looks like this is the public repository of all lightcurves recorded by Kepler: https://archive.stsci.edu/kepler/publiclightcurves.html https://archive.stsci.edu/kepler/publiclightcurves.html
    - cyphar 11y ago
      That's the first Kepler mission (which lasted 4 years, after which two of the reaction wheels used for fine pointing failed). There's also the K2 mission (which is a resurrection after the reaction wheel failure), which has much more interesting stars from an Asteroseismic point of view: https://archive.stsci.edu/k2/download_options.html https://archive.stsci.edu/k2/download_options.html You can also get the raw pixel data if you're doing some more esoteric data analysis.
- GFK_of_xmaspast 11y ago
  I think this was a case of researchers going back and analyzing data to look for a particular event, as opposed to "they've been crunching through the stream for years and are only now up to 2011".
- Sanddancer 11y ago
  No, astronomy keeps insane amounts of datasets, because things turn up often enough that what was interpreted as a star fifty years ago is actually a planet. Going through, looking for precovery images means that the more data you have, the more you can sift through later with more knowledge. The old data may never be analyzed, but it still waits there until we discover more about the universe to find its significance.
- kjs3 11y ago
  This is not unique to NASA, or indeed digital data. An anthropologist acquaintance of mine recently made a significant (though very niche) discovery sifting through samples collected on a dig done in the early 1900s. There's often not enough eyes to look through all the data out there as it's collected, and new eyes often spot new things.
losvedir 11y ago
Worth pointing out that the cool gif and video are just illustrations. I'm not sure what was actually "captured" when it happened. Maybe the chart on the side of the page which shows the spike in brightness?
- danparsonson 11y ago
  Yes the chart is what they captured; the animation is inferred from the data and our current models of supernovae.
  - mkagenius 11y ago
    That breaks my heart :(
    - imglorp 11y ago
      Recently we have started to see actual pictures (ie, not a point) of extrasolar stars. The keyword when searching is direct imaging. https://duckduckgo.com/?q=direct+imaging+betelgeuse&t=ffsb&iax=1&ia=images https://duckduckgo.com/?q=direct+imaging+betelgeuse&t=ffsb&i...
      - danparsonson 11y ago
        Indeed, see also https://en.wikipedia.org/wiki/HL_Tauri https://en.wikipedia.org/wiki/HL_Tauri and http://www.space.com/31497-exoplanets-direct-imaging-next-big-thing.html http://www.space.com/31497-exoplanets-direct-imaging-next-bi...
jkot 11y ago
It is a quite impressive. Kepler is pointed towards Milky Way disc and those stars are in other galaxy.

If I understand correctly, this data was recorded in 2011. It amazes me that it could take NASA this long to sift through its own (massive) data. As datasets are getting ever larger (outpacing gains in computing power, I feel) I wonder if some similar discoveries will take decades (centuries?) in the future. Does anybody know if these teams make it a point to keep datasets "small enough" to be manageable in reasonable time frames?

cyphar 11y ago
It should be noted that NASA makes essentially all of it's data public. In the astrophysics community, this means that researchers write scripts to run over all of the data for exploratory research and then do more complicated analysis later. It's also important to note that one of the best and most accurate space telescopes ever made (Kepler) was keeping the entire photometric community very busy.
- ezequiel-garzon 11y ago
  Thanks for your comment! Do you feel that, in the medium term, as ever larger datasets are generated the overall computing power will be unable to catch up?
  - greglindahl 11y ago
    Data gets bigger because instruments have bigger/better sensors (the biggest of which are built like computers) and there is more storage and bandwidth. So it basically advances in lockstep with computing power. There's a lot of hand-waving and details, but that's the overall answer.
  - cyphar 11y ago
    To be quite frank, the limiting factor is not computing power, it's code written by physicists. I'm a software developer by trade, and it's sometimes painful to see the type of code my supervisors write. I get the feeling that "our simulation takes half an hour" is acceptable and they won't try to optimise it unless they're near a deadline. Obviously, that's not deriding them, they have much better things to do than to micro-optimise their code. But in general I prefer writing all of my analysis scripts from scratch after discussing what methods they used. :P
- bartman 11y ago
  Looks like this is the public repository of all lightcurves recorded by Kepler: https://archive.stsci.edu/kepler/publiclightcurves.html https://archive.stsci.edu/kepler/publiclightcurves.html
  - cyphar 11y ago
    That's the first Kepler mission (which lasted 4 years, after which two of the reaction wheels used for fine pointing failed). There's also the K2 mission (which is a resurrection after the reaction wheel failure), which has much more interesting stars from an Asteroseismic point of view: https://archive.stsci.edu/k2/download_options.html https://archive.stsci.edu/k2/download_options.html You can also get the raw pixel data if you're doing some more esoteric data analysis.
GFK_of_xmaspast 11y ago
I think this was a case of researchers going back and analyzing data to look for a particular event, as opposed to "they've been crunching through the stream for years and are only now up to 2011".
Sanddancer 11y ago
No, astronomy keeps insane amounts of datasets, because things turn up often enough that what was interpreted as a star fifty years ago is actually a planet. Going through, looking for precovery images means that the more data you have, the more you can sift through later with more knowledge. The old data may never be analyzed, but it still waits there until we discover more about the universe to find its significance.
kjs3 11y ago
This is not unique to NASA, or indeed digital data. An anthropologist acquaintance of mine recently made a significant (though very niche) discovery sifting through samples collected on a dig done in the early 1900s. There's often not enough eyes to look through all the data out there as it's collected, and new eyes often spot new things.

losvedir 11y ago

Worth pointing out that the cool gif and video are just illustrations. I'm not sure what was actually "captured" when it happened. Maybe the chart on the side of the page which shows the spike in brightness?

danparsonson 11y ago
Yes the chart is what they captured; the animation is inferred from the data and our current models of supernovae.
- mkagenius 11y ago
  That breaks my heart :(
  - imglorp 11y ago
    Recently we have started to see actual pictures (ie, not a point) of extrasolar stars. The keyword when searching is direct imaging. https://duckduckgo.com/?q=direct+imaging+betelgeuse&t=ffsb&iax=1&ia=images https://duckduckgo.com/?q=direct+imaging+betelgeuse&t=ffsb&i...
    - danparsonson 11y ago
      Indeed, see also https://en.wikipedia.org/wiki/HL_Tauri https://en.wikipedia.org/wiki/HL_Tauri and http://www.space.com/31497-exoplanets-direct-imaging-next-big-thing.html http://www.space.com/31497-exoplanets-direct-imaging-next-bi...

jkot 11y ago

It is a quite impressive. Kepler is pointed towards Milky Way disc and those stars are in other galaxy.