I was curious if anyone who worked on the new Enhanced Light Curve Generator thought of possibly adding the ability to screen outliers. Following Betelgeuse, I was studying the trend and means and thought that it would be nice to have a button, in addition to the mean, a "median" button too. That would inherently screen out the outliers unlike the mean and provide, IMO, a more relevant mid-point.
Another more cumbersome option would give the user the ability to click data points and remove them. Still another would be to set a max and min preference for data points. Currently, there is an option to set min and max in preferences but that is for the axis scale; it doesn't remove the outliers but is very helpful in visualizing the trend. A median button seems like a simple, best option.
Re Betelgeuse, the latest from the fine folks at Astronomer's Telegram noted that they recorded the minimum magnitude at 1.614 in the 7-13 Feb period. Our LCG shows a mean of 1.602 for those seven days but included data points as low as 2.0+ and as high as 1.2. Because enough data points were submitted, the 7 day period mean with the LCG was near equal to what Astronomer's Telegram reported.
For smaller day ranges, and for stars where the submissions are not as heavy, some outliers will have a greater impact. If this were back in the day before photometry and other hi-tech measurers were available and only visual estimates were available, scientists would have used a median or other ways to get the best possible mid-point measure.
This is just one of those "would be nice things". The Enhanced LCG is a very nice tool as is.
Update Edit: Out working in yard and realized there is a list of contributors if you scroll down and likely will be able to find the data points in question to uncheck. Sorry. That's what being new does sometimes.
Cheers, Bob
Hello Bob,
You might want to take a look at VStar if you are interested in analyzing data (https://www.aavso.org/vstar). This tool gives you a lot more flexibility in what data to display using filters, and you can click on outliers to remove them from your plot. It will also let you plot a mean curve (you can set the bin size). Please be sure to read the VStar User's Manual which explains its many features.
LCGv2 was really meant to be a quick way to plot data and see a light curve with a minimum of features.
Best regards,
Sara
Thank you for that suggestion. Being new, I saw reference to VStar but it wasn't one of the features I got around to yet. Nice to hear it is more sophisticated. The LCG I found to be very nice too (especially since I discoved the contributors/data points at the bottom can be selected to include). I will be reading the VStar manual as next project. Thx!
Bob