I am actually going to be spending my summer improving this project at the Time Series Center, and also making a better API for it so that I can give it to some other people who have asked about it and maybe even release it online. One of my big goals is to allow users to search for similar lightcurves and be able to scroll through a list of matches. Also, I would like to allow users to compare multiple light curves from different objects. I also want to add in some additional features, like folding the lightcurve around its period or highlighting astronomical events in the data. I also am a bit unhappy with the "Lightcurve Library" bar at the top, as it is not perfectly straight-forward what it means, so I might re-design it. Also, I want to add some ability for users to view the raw, un-binned data. Finally, I originally wanted to allow for linking and brushing between the two elements, so that as the user moves his or her mouse over the points on the Google Sky map, they would quickly load into the graph. However, the load time is a bit too slow for that sort of feature, and so I opted to just have users click the points instead to load them.
The main purpose of this visualization is to allow astronomers to browse the vast quantities of data available to them, find trends or events, and check reliability of the data.
Also, I wanted to be able to find similarities and differences between different lightcurves, both those for the same object recorded at different times, and those from different objects entirely.
Some questions:
This visualization proves that the lightcurve data can often have errors in it, but by showing a lightcurve along with other lightcurves, these differences stand out and are easily noticeable. This helps astronomers to make sure the data they are viewing is reliable and to not treat an error in the data as an astronomical event.
The lightcurves also show how similar objects are in the sky. The user can easily view different stars in the same region by clicking on them in the Google Sky map.
The ability to plot multiple lightcurves at once shows how lightcurves taken at the same time but on different channels tend to follow the same overall trends.
One of the unexpected insights we made about the data is how sometimes lightcurves of the same start may have been recorded but not linked together, and therefore they are listed as separate (but nearby) stars. By viewing the lightcurves in this graph, you can notice the similarities between two lightcurves and figure out that the two different datasets both correspond to the same star.