LAB NOTEBOOK

Final Post

Jun 4, 2021

I'm graduating tomorrow! What a journey this has been.

The project isn't quite done yet but we've made significant progress since the end of Q3 and are in the end-ish stages of our paper currently. The set of slides I presented for the Wildcat Colloquium and for my senior passion project can be found here.

Final Quarter 3 Reflection

Mar 27, 2021

This was a big quarter for both our group and our project—I am quite happy with and excited by the progress we've made over the past few months.

Since the start of Q3, we've completed both our analyses of actual TCEs and our experiment to assess our reliability. Now, as we continue to assess our performance in the aforementioned TCE analysis, we're going to be starting the paper that will describe and present our work over the past several months.

That, I will admit, feels quite strange to write. I don't think I would have expected back in September to be working with my group on a paper with the intention to publish. I'm immensely proud of that fact and of where we as a group have gotten in this project, and I feel extremely grateful to have had the opportunity to be a part of the MARC program, to have Steve as my amazing mentor, and to have had Stori as an incredible MARC instructor. Having just had our last official MARC class, I know I'll miss meeting weekly as an entire cohort. Nonetheless, I am definitely looking forward to continuing with the Kepler project and after all of our group's hard work, to having a paper and a final presentation to show for our efforts.

Progress Presentation Reflection and Updates

Mar 26, 2021

Although the Q3 progress presentations came at a somewhat stressful time for me, I am quite glad after the fact to have put it together and presented to my MARC classmates.

Firstly, in the process of assembling my slideshow, I was prompted to return to some of the literature we've referenced previously in earlier stages of our work with Kepler data. Of particular note was the paper I wrote about in the previous post. It was actually quite helpful to re-read these resources, as doing so helped to solidify my understanding of how Kepler data was processed and then organized. I was also quite interested to be reminded of the parallels between our own study's high-level design and that of previous ones conducted using Kepler data.

I will admit that I was quite short on time to practice for my presentation after putting together the slides. However, rather than feeling under-prepared, I actually felt as though I was able to give a natural and smooth presentation. Doing so actually made me realize that I am able to fairly comfortably explain and talk about what I've been doing with Kepler data over the past year and a half. This was quite a satisfying realization to have—I suppose it's often hard to grasp your own continual progress over a long period of time.

Paper Showcase

Mar 22, 2021

One paper that I've spent a fair bit of time with recently as I've been putting together my presentation is "Planetary Candidates Observed by Kepler. VIII. A Fully Automated Catalog with Measured Completeness and Reliability Based on Data Release 25" (Thompson et al. 2018).

The paper, in summary, presents the DR25 Kepler Object of Interest (KOI) catalog, which was the catalog resulting from the final Kepler data release. The DR25 catalog contains 8054 KOIs, 4034 of which are planet candidates (at the time of the paper's release) (Thompson et al. 2018).

Of particular relevance in this paper is Robovetter, the automated system that produced the DR25 catalog of KOIs by vetting and sorting TCEs (any observed repeating signal in Kepler data) into planet candidate (PC) and false positive dispositions. Each TCE was assigned a "confidence" score ranging from 0 to 1, with higher scores indicating higher confidence that a TCE is a PC. Only transit-like TCEs—those that are likely to be either transiting exoplanets or eclipsing binary stars—were included in the KOI catalog (Thompson et al. 2018).

This paper is particularly relevant to our project because we are currently looking at TCEs that may be planet-like but due to low statistical confidence, as determined by Robovetter, did not end up in the KOI catalog. In fact, our group is actually somewhat mirroring how Robovetter was used in the process of producing the DR25 catalog, except by hand, instead of an automated system.

First, we completed our analysis of TCEs that met our criteria, using similar metrics to Robovetter to determine tentative dispositions for each of said TCEs. Then, very similarly (in a general overall sense) to the process in Thompson et al. 2018, we performed the same analysis we initially conducted for our real TCEs, this time on simulated TCEs, to assess our own reliability in vetting TCEs ourselves.

We hope to, in doing this work, contribute in some way to the continuing search for exoplanets.

Project Updates #7

Mar 19, 2021

As of Monday 3/15, our group has officially completed the "TCE analysis" phase of the project. Steve, Halle, and I met together with Jeff to finally learn how we performed in the validation experiment, which marks the conclusion of our experimentation.

We chose to have Jeff tell us in real-time what each of the "fake" TCEs we analyzed were—injected, inverted, or scrambled. As it turned out, this was a pretty suspenseful, but satisfying, process. We did end up achieving a reasonable—although not perfect (which would be quite unrealistic)—level of accuracy in our analysis of said simulated TCEs. After getting "scored," we began putting together a table to quantify our performance along various metrics pertaining to the type/disposition of TCEs, etc.

However, at the moment, it's hard to say specifically how accurate we were, as we have yet to finish analyzing and taking statistics for the experiment as a whole. Next steps for our group will entail completing that meta-analysis, and then officially beginning the writing process—I'm very excited to see where this goes.

Project Updates (Senior Passion Projects...)

Mar 13, 2021

Amidst further progress in our project and new announcements from our senior class dean, one thing I've been thinking more about lately has been how senior passion projects relate to MARC.

At this point in time, it seems quite likely that the writing phase of the project (due to start Monday 3/22) will last until the end of the semester. Given that, it would be reasonable to assume that the two weeks of passion project work time will be spent working on the final parts of the paper and on wrapping everything up.

However, as part of the passion project there is also a presentation component, to be given to members of the greater MA community, separate from the official MARC- and Kepler-related final products. As such, if it ends up being feasible within time constraints, I think it could be quite interesting to explore different modes of presenting what I've spent the last year and a half doing in MARC. One thing that comes to mind was the "Dance Your PhD" video that we saw during class last year. I'm not a dancer, so I couldn't do exactly that for my project, but a similarly creative approach might both allow me to demonstrate deep understanding of my topic and project and to possibly present more effectively and in a more engaging way to a more general audience.

Updates on this particular subject are likely to follow later on in the semester.

Project Update #5

Mar 5, 2021

Since the last blog post, Halle and I have finished with all of our initial analyses of the TCE sets. We've made steady progress analyzing those during our meetings with Steve—we'll need probably one more meeting to finish completely with that phase of the project. Most likely this will happen at the next meeting, which is scheduled for this Monday. At meeting following that one, as I have mentioned in the past, the accuracy of our analysis will be scored by Jeff.

In the meantime, the writing process for the paper has also begun. Halle and I will contribute to the methods section, which we'll start working on with Steve very soon during a one-off writing session that we will have in addition to our weekly meetings. I'm very much looking forward to moving into this next phase of the project, as I'm eager to exercise some of the skills I've practiced throughout our MARC coursework and to gain more "official" real-world experience working on a full-fledged scientific paper that will ultimately be submitted to a journal for publication!

The next post will probably not be a progress post per se—I'm looking to explore some other types of content to add some variety to this page.

Week 4 – Project Progress

Feb 23, 2021

Over the past week or so, we've been making good progress with our analysis of the simulated TCE data. I met with Halle over the weekend to complete our initial assessment of the second set, and then we all met with Steve as a group to collaboratively review and analyze it. On the whole I think (I hope) that we're doing a pretty accurate job so far. It's definitely paid off to have done the same type of analysis on so many TCEs by this point—my understanding of the process (e.g. what to look for, how to make decisions with regard to the disposition of the object) continues to grow deeper and more intuitive. However, it has been a bit strange to have to approach these simulated TCEs as though they're real, observed objects while knowing that they are in fact not so.

We expect to take a couple more weeks to complete our analysis of these objects, at which point we'll get "scored" by Jeff on the accuracy of our methods of analysis. At that point, we can then move into full-on writing mode for the paper that will describe our work over the past several months.

I've also started to think about the methods section of the paper in particular, and have put together a preliminary outline, which has helped me to get a sense of the content and structure it might have. Steve is also in the process of starting to write the paper as well. I'm hoping to have some discussion soon about how the writing process will work and how I as a student might contribute—I'm very excited to see where things go from here!

Week 3 – Project Progress

Feb 15, 2021

A lot has happened with our project since my last blog post.

First, in a meeting on Feb. 1, we—Steve, Halle, Jeff (a colleague of Steve's), and I—were able to review the TCEs that we think may be potentially planet-like. Jeff agreed with our initial assessment, which was a promising step forward for our project. During that meeting and the next (sans Jeff), we also reviewed the design of the next phase of this project.

Essentially, without going into too much detail, we're currently in the beginning stages of a double-blind experiment using simulated TCE data with known dispositions (but unknown to us), randomly gathered by Jeff, to help "validate" our analysis of the real TCEs that we previously examined. Halle and I met earlier this week to begin our experimental work, collaboratively reviewing the first subset of these new TCEs. We'll bring these to the table at our next meeting later today for a review session with Steve. Then, we'll repeat that process over the next few weeks until we're done with the set of simulated TCEs.

Things have been going pretty smoothly for us with the project lately. However, in the past couple weeks I have run into some technical issues with Jupyter, Python, and some of the packages we've used for our analysis this semester; I've been working to resolve these with Steve's help. Our current hypothesis is that these issues are resulting from incompatibilities between my new computer's architecture and a few packages and their dependencies. While they're not totally resolved yet, things seem to be almost working—but not quite.

Luckily, this isn't hindering our work at the moment, and solving this puzzle has actually ended up being somewhat interesting and fun for me.

Beginning of Quarter 3 Updates

Jan 27, 2021

I'd like to use this space to provide some updates on the status of the Kepler TCE analysis project as Q3 gets under way.

For Q3, one of our main goals as a group is to continue our deeper analysis of TCEs (via examining more closely the light curves of each one), having assembled, through our earlier phases of investigation using the information provided in the TCERT reports, a list of TCEs that we think are potentially planet-like enough to be worth looking at further. As this phase of the project begins in earnest, we will also need to begin thinking and discussing as a group about what the endgame for this project will be, as far as the "final product" at the end of the semester. No matter what, that will at least include a presentation of our work at the year-ending Wildcat Colloquium.

I went over these goals for Q3 with Stori in our 1-on-1 meeting on Monday, and talked about the plan for my group's weekly meetings (Steve, Halle, and I are about to start up with a regular meeting for the quarter). We also talked about how we'll be spending MARC class time this quarter—the writing workshops that we talked about will prove particularly useful for where this project is hopefully going over the next semester.

Most of the potential roadblocks I'm concerned about are related to project planning and time management, but I think these will be possible to avoid if I communicate effectively with my mentors and group-mates.

Quarter 1 Reflection

Oct 16, 2020

This blog post comes at the end of Q1 of my senior year.

I've enjoyed the work I've been doing since the beginning of the school year (or realistically since before then, considering the summer work). It has been good to get to know and to work with two new members in our research group.

On the whole I've been quite impressed with our group's speed in moving through the preliminary TCE categorization work. Our weekly meeting schedule, both amongst the three of us students and also with Dr. Steve Bryson, has been effective in this regard, as it has seemed to be the ideal amount of time to go over each week's work. Ditto the weekly class meetings with our MARC cohort.

As far as MARC course materials, I think that what we've done this semester has been helpful. Particularly, the Excel and writing workshops were helpful. The Excel workshop was good in that it led me to think critically about the tools I'm using in my research, even if I didn't end up using Excel (I decided that Pandas, MatPlotLib, etc. are better for my purposes). The writing workshops were good for getting me to start thinking about what I ultimately want to produce for this project, and about how that will begin to take shape in the coming months.

As I've outlined earlier, my goals for Q2 are to continue the great research work we've been doing thus far—preliminary analysis of the list of TCEs we're working will be done relatively soon, and then we can move into more specific analysis of the ones that are of note. Tentatively, too, I plan to start working on a methods section for my project, though depending on my schedule this may actually happen in Q3 instead.

Quick Project Updates

Oct 10, 2020

I think this will be a shorter blog post for this week, as I've been quite busy for reasons that do not relate to my MARC project.

This week my goals were to:

Continue reviewing our sets of TCEs
Meet as a whole group to review said TCEs

Our first meeting with these objects went well. As a group, we've made a lot of progress working through the TCEs that we've set out to look at.

Essentially, these TCEs are potentially transit-like events initially flagged in the Kepler data. The subset of TCEs we are evaluating are those that have not been included in the KOI database, and fall within a certain range of (very low) confidence scores. Among these, we have already found several potentially interesting candidates for further examination, ranging from potential planet-like objects to binary star systems, etc. We will continue to move through our list of TCEs, and then will likely shift over to analyzing the individual ones that we have decided are of particular interest.

Despite the scheduling circumstances of next quarter, which are outlined in the previous blog post, I'm excited to continue working on our project, as I think we are going to make a lot of very exciting progress over the next few months.

Project Updates – Quarter 2 Goals

Oct 3, 2020

This week I've spent some time working on establishing SMART goals for Quarter Two.

It'll be a bit different from this first quarter in that our MARC IR2 cohort will not be meeting on a weekly basis as we have been. Other than the specific workshops we've been doing and being able to discuss various things as a group, the lack of a class time won't have that much of a direct effect on my project. My (weekly?) meetings with my own research group will continue to happen, and I think we'll be able to make good time on the TCEs that we have set out to evaluate, which we have just started doing.

In some ways, I think it'll actually be easier from a logistical perspective to only have one set of meetings to plan around, because during the first quarter it has been difficult at times to manage my project around the Monday weekly meeting schedule, because my own weekly project schedule is offset from that by about three days.

My goals are essentially to keep moving the project forward as we have been. This will include various sub-tasks:

Keep doing the actual object analysis.
Meet as a group on a regular basis, both independently and with Dr. Steve Bryson.
Work on a methods section to sort of codify in writing what we've been doing.

I think I will be able to get the support I need, as my group will continue to meet, and I can always meet with Stori as well.

Finally, I've received notice in the past few days that we'll be returning to campus for at least part of the upcoming quarter. This does not affect the project much in itself, but it will present some scheduling challenges. We'll see how this pans out, but I am optimistic that we will be able to make good progress during this time.

Project Updates – Jupyter edition

Sept 28, 2020

I wanted to take the time this week to do a bit more of an informative post, rather than just another project update, though my project is moving along nicely; we are starting to get into our research question and related topics.

Specifically, this post will be about our software setup with which we can manipulate, analyze, and visualize data. The backbone of this is the Jupyter notebook, which allows us to run Python code in real-time, with each output displayed right below the code that generated it. Below is an example of a Jupyter notebook:

The libraries we are using in this example notebook (though others are also imported) are Pandas, which allows data to be manipulated and analyzed in the form of tables, and MatPlotLib, which is a powerful graphing tool. Here, we have an example table of four different types of animals, with various numbers attached including the number present, their size, and their weight. It's quite easy and useful to select only part of a data set; a new table can be made containing only the desired data. If we wanted, for example, to only select data for animals larger than 0.5 Smoots, then we could do the following:

Of course, while it is useful to be able to manipulate data in this fashion, it is also useful to be able to visualize them. This is where MatPlotLib comes in. For example, if we wanted to create a scatter plot of animal size vs weight, it is quite simple to do:

Example bar chart representing the number present of each animal:

I've found these libraries, amongst others, to be very useful in working with the Kepler data. Often, I've found that it's actually quicker and easier to use Jupyter notebooks than Microsoft Excel when working with large data sets in this way, not to mention requiring less of my computer's resources.

Project Updates

Sept 19, 2020

We had our meeting as a group, in which we reviewed our progress with the KOIs that we had previously decided to work on for the week. I am happy with how this is going; I think we're transitioning nicely into the more specific project work. We currently have a new, more difficult set of objects with which to practice, which are more in line with the type of work we're going to be doing; I plan to meet with Talia and Halle soon to work on this collaboratively.

I also spent some time organizing materials so that I can put together a writeup of my methods section for the project, though I haven't started writing yet. I collected past descriptions I had written, as well as the methods section from the poster (which contains the most well-defined process of anything I produced last year).

Project Updates

Sept 12, 2020

I met with Halle and Talia to review the designated objects; it was a productive session and I enjoyed being able to work on the investigative aspects of this project after a long period of preparation and organization.

I also spent some time using Excel to work with the data. I do, however, still feel more comfortable manipulating and graphically representing data with various Python libraries.

Finally, I prepared for and attended the guest speaker event, in which two researchers spoke about their work and stories so far. I found the session informative and interesting, as their fields (psychiatry/drug therapy and virology (particularly in marine mammals)) are quite different from the one in which I currently research. In addition to the actual material of their research, it is always good to hear from established scientists how they have arrived at their current positions and projects.

Project Updates

Sept 6, 2020

We met as a group on Thursday to discuss a number of topics and continue our preparation before we dive into research. Such topics included version control procedures, file/repository management, and continued work with example KOIs. As a group, we practiced analyzing several KOIs with different characteristics, including both planets and false positives (e.g. background binaries, eclipsing binaries, etc.). Our group will work together on a set of "practice" objects for homework before our meeting next week.

On the other side, I spent time working on logistical and organizational work for the project, including organizing my resources that will be needed. I also began a document that details goals for each week and summarizes what was accomplished.

Start of School Update – Fall 2020

Aug 30, 2020

I have been working with Dr. Steve Bryson as well as two new members of our group over the summer to get up to speed with and review material that will be relevant to our project this coming year. We are now beginning our transition to project-focused meetings, in which we will discuss and analyze work that we have completed on a weekly basis. The first meeting will likely be in this coming week.

I thought the readings assigned by Stori were informative and helpful. I thought the emphasis on holding yourself accountable for your own productivity was good. All of the tips and methods of planning and organizing are good, but a plan, in my opinion, is only as good as its execution. Therefore, it makes sense to go a step further and implement methods such as an accountability partner etc so that you stay on track and actually adhere to your plans, meet your goals, and finish what you start. Also, I thought the specific organizational tips were helpful, especially those suggesting version control and repositories for computer-stored materials—in fact, we are already using a GitHub repository to store and share our files as a group.

My SMART goals for this month are to get situated and begin investigating the Kepler objects which we will outline in our upcoming first meeting of the school year. I think these are doable, if a bit conservative. I will re-assess after our first meeting so that I can be more specific.

Final Reflection – Spring 2020

May 26, 2020

Had I been told back in September what I had done this year in MARC, I'm not sure I would have believed it. I truly did not expect to have found an incredible mentor so early, been part of an amazing group, and actually had some results. Looking back, I can't help but realize that a lot of this has come from sheer dumb luck, just because had I not approached Steve on a whim at the panel, I would not be where I am today.

Our group has tentatively put together a list of results for our inquiry. These are then to be published pending examination by the False Positive Working Group. I'm extremely happy with my poster and proud of all the work I put into it, as I feel it embodies everything we have done this year. I feel that it effectively presents our research, and looks good in the process—it's something that I can look at as sort of a symbol of what we have done this year.

It's not entirely clear what path we will choose to take from here, whether that be continuing our line of work with Kepler or tackling some other, equally fascinating astronomy project. However, no matter what this summer and next year hold for the project, I look forward to seeing what the future holds and am eager to continue working!

Project Updates

May 2, 2020

In our last meeting as a group, we continued to review certain objects that we had selected as perhaps having been of importance. The list of these was of medium length, and certain objects were identified that had characteristics that may necessitate further review. One of these was initially thought to be a particularly interesting and rare type of object, but the phenomena that indicated this were later shown to be artifacts of the light curve signal.

Project Updates

April 17, 2020

Our group chose to take a brief hiatus from our work that roughly coincided with my school's spring break. Thus, progress has been limited since our last meeting, during which we were able to begin to glean patterns and interesting bits of information from the new groups of objects that have been obtained in various ways. While my mind has not wandered too far from our research, I am also enjoying having more time to relax, be outside (> 6 feet away from everyone else). Next steps will be to continue investigating the most recent objects that have been identified as noteworthy, and ultimately perhaps further pursue some patterns that have already started to emerge. Also, I recently read news about a new planet discovered from archived Kepler data that orbits in the habitable zone of its star (the article can be found here). This is particularly motivating and exciting for me because the general idea is quite similar to what I am currently doing, and shows that there are indeed many possibilities for new discoveries to be made from the very extensive archived data from the Kepler mission.

Project Updates

March 19, 2020

In a recent meeting with Dr Steve Bryson, Nick, and Aurelia, a new potential direction of investigation and analysis was determined. Based on an object that was previously identified, we were able to search for a specific set of similar objects. This was a new method of investigation for us; it will provide a way of gaining more specific access to certain sets of objects with particular characteristics. These objects may be significant for a number of reasons pertaining to the way false positives were detected and vetted in the original Kepler mission, and their investigation may yield a better understanding of the systems that carried out this detection and vetting. In doing so, we may help to further the search for exoplanets and contribute, if somewhat indirectly, to the creation of more precise and efficient systems of separating out false positive objects from planet candidates, and to decrease the likelihood that an object which is recorded as one type actually being of the other. Ultimately by refining such systems we will gain a better understanding of the number of planets within our galaxy, as well as how these are distributed amongst types of stars and perhaps further down the line, when a more global survey is possible, how the numbers of planets may change regionally throughout the galaxy.

Presentation

March 3, 2020

I found my recent progress presentation to be an informative and fun experience. I greatly enjoyed being able to share the findings of my research thus far. However, it was a challenge to fit all the requisite information within the time limit, and it was somewhat difficult to prepare—I found that for me, although preparation is absolutely necessary, it was be more helpful to loosely plan out the points that I needed to hit and then fill in the more nuanced aspects as I presented; this allowed me to feel more natural and organic as I gave the talk. Also, I took great satisfaction in being able to interact with the rest of the program, both as a presenter answering audience questions after the presentation, and as a member of the audience myself; I enjoyed learning what my peers have been doing in their research as well.

Progress

February 13, 2020

I recently met with my mentor, Dr. Steve Bryson, as well as two other students with whom I am working at the moment. It was refreshing to be able to discuss the project in-depth for the first time after a month or so of not doing so. During this meeting, we spent a lot of time looking through some of our findings within the Kepler data. It was ultimately determined that the aspects of the data that we have thus far looked at may be a viable direction of research. Also, we discussed taking the project in a different, but related and very interesting direction. I look forward to presenting what has been found so far and to our next meeting.

February Goals

February 5, 2020

I have continued to examine the objects that have decreased in size. Some decreases may seem significant, but the next step is to bring these to the table during my next meeting with my mentor, Dr. Bryson, on the 12th. My goal is to determine whether or not this is a viable research direction at this meeting. If this ends up being the case, I will then begin examining these objects methodically to identify any potential patterns or particularly significant objects. I am excited to be able to begin preliminary investigation for my project, but I of course need a more experienced opinion to decide whether or not this particular aspect of the data is worth looking at.

Other goals are to revise my resumé, which will appear on this site soon.

The Gaia spacecraft.

(Image source: https://commons.wikimedia.org/wiki/File:GAIA_(14050944939).jpg)

Revised KOI Radii from Berger et al. 2018

January 23, 2020

Berger et al. 2018 has the potential to change a number of KOI radii. Using the ESA's Gaia spacecraft, much more accurate radius data was collected for Kepler stars than was available previously. Because the depth of a given transit is proportional to the radii of the host star and transiting object, any change to the radius of a star leads to a resulting change in the measured radius of the orbiting object. While the majority of the- objects in the false positive table I'm working with experienced some change, it is the ones which cross the 30 earth-radius line that I'm most interested in, as that is considered the cutoff when an object becomes too big to be a planet. Currently there appear to be four KOIs which grew from less than 30 earth radii to greater than 30 earth radii when using data from Berger et al. 2018, while 43 decreased to below that threshold.

Further analysis will of course be required, but it is tentatively possible that these changes in radius may merit the reclassification of some false positives to possible planets and vice versa.

Report abuse