The survey's observing record, computed from survey.exposures. These
are ZTF's observing statistics, not Visnjan-Washington's — the stand-in notice above
applies to every plot here as much as it does to a search result, and the synthetic
fixture is excluded by default because its cadence is invented (a perfectly uniform
one exposure every 0.01 day), not observed.
How often the survey observed, split by band. A gap
is a week it didn't run, not one with nothing to see.
Hover a column for its date, its total and how that total divides
between bands, or click a band in the legend
to drop it from the chart.
Zoomed to
Jan 1, 2019
– Dec 31, 2019;
the from and to in the address take any date, not just
the whole years the chips offer.
The same data, accumulated — flat stretches are when the survey was not running, which the bar chart above makes easy to miss among the weeks that did have data. Hover a point for the running total and what it added.
How the observing above divides between bands — always the whole split, whichever band is currently picked. Click a slice to show only that band in the charts above, or click the one already picked to come back to all of them.
How often the survey comes back to the same object, over the same bands and
dates as the charts above. Grouped by ZTF's own association of the stand-in
data (ztf_oid), which is the only thing linking one detection to
another until phase 2 builds objects — so these describe what the stand-in
catalog can support, not a VW data product.
Only the 2,279 detections ZTF associated this
way are counted, out of 1,002,279 sources in the
catalog: a source with no association has no object identity here at all, so
it is absent from these charts rather than appearing as a one-point curve.
With a band picked, a gap is the time to the next
detection in that band — which is the interval that matters when you
fold a light curve.
The sampling, which is what sets the periods a light curve can recover: a signal shorter than the typical gap is not resolved, and one near a spike in this histogram aliases against it. Log-spaced buckets — revisits run from minutes to years. Hover a row for its share.
How long the curves are. An object with a handful of points has a position and little else — fitting a period or an orbit takes tens of points at least, so it is the bottom rows here that are worth plotting. Plot the best-sampled one (13 points).
Same data as JSON: /api/statistics?band=zi&from=2019-01-01&to=2019-12-31