Research Project
My particular project involved using IDL to closely examine data retrieved by Steve Howell and Mark Everett from the Planetary Science Institute. Without specific instructions on what to set about doing, I decided to look at the data and try to find any patterns, correlations, or problems with the statistics and information that I was given. Meanwhile, I learned a good deal about the IDL programming language.
The information I was given included apparent color magnitudes, positions, variability, best found periods, amplitudes, and statistical errors. The data was taken over five nights for almost half a night each. I was sent plots of the apparent V-magnitude over time (in Julian days) of those stars that were determined to be variable using statistical analysis. All of the information was divided into eight ccds.
Before I had access to the data, I attempted to classify the plots by simply looking at them. This was not very successful or accurate; so when I had the data accessible through IDL, I spent time making lightcurves (V-magnitude versus phase) using the periods given. The phase was determined by dividing the change in days by the period. Once I had the lightcurves of all of the tagged variable stars, it became apparent that several of the curves looked peculiar. This led to an interest in examining the periods themselves.
In the early stages, I had decided only to look at the variable stars from ccd 1 in order to reduce confusion. After reading the periods into a program and sorting them, one of the periods seemed odd. While most of the periods were between less than half a day to less than ten days, one period was 66.888672 days. After including the data from the other ccds, I found that several variable stars had been assigned the exact same period. It seems rather impossible for any two stars to have the same period out to six decimal places.
List of Duplicate Periods:
Frequency of the Periods:
third column: period (in tenths of a day)
fourth column: amplitude
fifth column: tag (used in other programs -- if the value is one, the period had
been duplicated.)
third column: frequency (notice that 55.740561 was repeated 12 times)
After noticing this problem, I used IDL to classify the stars by period. I also analyzed the
periods themselves and found that there were several duplicated periods. In fact, after
excluding the stars with duplicated periods, only 91 tagged variable stars were left out of the
original 193.
Reduced List of Variable Stars:
third column: period (in tenths of a day)
fourth column: amplitude
fifth column: average V-magnitude
Meanwhile, I had been in the process of plotting the colors of the stars versus period and
amplitude to check for anything interesting. I eventually excluded the stars with duplicated
periods from the plots even though some of the stars with very well-defined lightcurves had
duplicated periods. If I had more time, I would re-include them; however, with
so few of them, I do not feel that they would have much impact on the plots anyway.
Below are links to the more interesting plots.
I also attempted to devise another classification scheme. Below, I have included a link to an explanation of and examples from each category. Many of the plots are difficult to classify, so the plots within the categories are always subject to revision.
A continuation to this project would be to include error bars with all of the plots, take spectra, and refine the classification scheme. The stars cannot be accurately classified without knowing their spectra, distances, and/or absolute magnitudes. A better period fitting program could also be beneficial. Unfortunately, I am out of time.
