Explore simple stats to run on each ZTF light curve

[jradavenport]

(e.g. 5-50-95% range) to select Boyajian's Star candidates. Check relevant literature on the idea (e.g. https://ui.adsabs.harvard.edu/abs/2019ApJ...880L...7S/abstract)

@dhuppenkothen wants to do some literature diving here

[dhuppenkothen]

🏊‍♀ (this is as close to a diver as I can come in emoji form)

[dhuppenkothen]

Search algorithm from paper above:

• omit all measurements that fail the “good” criterion based on processing flags (are there flags like this for ZTF?)
  • remove all measurements that have listed magnitude errors >0.1 mag
  • treat back-to-back pairs as independent observations
  • for each star:
    • compute median magnitude and standard deeviation from all of the magnitudes,
    • then trim the sample at three standard deviations and
    • recompute median and standard deviation from the trimmed sample: these form the continuum magnitude m_cont and the continuum standard deviation, sigma_cont
    • search for points more than 3sigma_cont fainter than m_cont
    • select all stars with at least one dip consisting of five or more contiguous low points no deeper than 0.75mag for visual inspection
    • visual inspection to reject:
      • any star with a light curve that resembles an eclipsing binary (i.e. dips wider than a few days, stars with regularly spaced minima)
      • any star with a light curve resembling an intrinsic variable
      • download finding charts from NASA/IPAC, reject stars with coordinates centred between two or more stars (i.e. blended sources)