I'd like to thank Nick Raymond for helping me out this week. He has asked lots of questions, which gives me an opportunity to describe and defend what it is I think I'm doing, and he read through my research paper with some really great advice. So thank you Nick!

I'd also like to thank Jen Valenzuela and Ellen Jacobs for being the best lesson plan guinea pigs any future teacher could wish for. They not only humored my lecture, but played along with the games and gave me some great advice. So thank you Jen and Ellen!

I have made some really great progress towards my 4 projects. Paper is almost finished (lots of revision left), Presentation is basically finished (need to input tables), Lesson Plan is finished, and my Poster is done, just need to make more pictures and take out some words. If everything goes according to plan, I will be right where I need to be by the beginning of next week STRESS FREE.

I sure am going to miss MBARI when this is all over

Juggling project stuff, presentation stuff, and paper stuff has been real fun this week...yes, I'm being sarcastic. Adjusting to the new housing has been a bit of a stressor (campus IT has yet to fix the internet issues at our house, yikes!). A recent passing of family member has also been weighing heavy (service is this weekend); it's been no easy feat maintaining clarity and focus, but my work's got to chug along.

I'm clocking in some solid time in Monterey this week...yesterday (Weds), Desmond and I conducted an informal survey outside of the aquarium. What better way to gather input from the general public than by interacting with them first hand! Though we both agreed it would've been better to have done this earlier on in the summer, the feedback we received gave us some great insights on the general public's understanding of our respective project topics and their information sources.

This morning, I returned to the aquarium's weekly story meeting for my marketing/social media/communications fix. I anticipated discussion over Instagram's Stories feature, which was unveiled this week and is identical to Snapchat. Like other businesses and organizations exploring the feature's functionality, the aquarium is strategically figuring out the next steps in utilizing it to their advantage. As pertains to MBARI's Instagram account (which turns one year old tomorrow), hopefully this new feature will encourage creating/sharing more video content and segue into other video-centric outlets, paralleling the growing popularity of video content in the realm of social media and public outreach. I also helped Dahpne, the aquarium's social media lead, in tackling an Instagram analytics issue that my mentor Susan had also encountered.

I have a few other meetings today with various aquarium folks and am hammering away on project/presentation/paper stuff in between, giving my eyeballs the occasional break with the overcast, but still gorgeous, views of Hopkins.

I think that everyone is stressed out about finishing their projects, presentations, and paper, but worry not! Everyone will do great

Jen and I conducted some informal surveys around the aquarium area yesterday. Much to my surprise, people were nicer than I had anticipated. Most people were willing to take our surveys and those who declined did so really nicely (for the most part). I spent today compiling the data, analyzing it, constructing bar graphs for the presentation, and determining whether a statistical analysis is necessary.

I am learning some really interesting things about the public but I won't spoil the fun here, you'll all find out next Wednesday.

My story map that I have been working on all summer has been given the thumbs up by all parties (researchers, industry professionals, and my mentor)and I need to write a statement to go along with the release on Monday or Tuesday.

Here I am Sunday night having seriously forgotten to do confluence, but I'm not the only one so I think that puts into perspective where we all are at with our stress levels. I did a practice presentation last Thursday and got some really good feedback from the lab group I presented to. Thanks to that I now only have a few tweaks left for my presentation! But alas it is not over yet. It kind of just hit me that I go back to school in two weeks which means while also having to complete this paper, I have to complete my actual thesis project, and figure out my life since this is my last semester.... Yay me! But it'll all work out the way it is supposed to! I hope everyone takes care of themselves this week because it's times of high stress when you can find yourself easily catching a cold or something.

I have been working so hard during this whole week that I even forgot to do confluence on time!
I think we're all more or less in the same position getting so stressed with the paper and the presentation but even though I am too, I'm happy with all the progress that I have made during this week. I made my presentation and I presented to my mentors, they told me that I made a really good job and that there's just a few things to change (most of about grammar and the way I explain things) so I can spend more time practicing for next Wednesday. I have an outline for my paper and a lot of comments from my mentor about how to make it better so I'm more stressed trying to write it than with my presentation, but I have a lot of people willing to help me with this yay!
Tomorrow I have a meeting with my lab group which again is a great way to practice for the symposium and also a good way to try to control my nerves when I talk in front a lot of people.

So I realized this week that the tidal harmonics results that I was getting simply could not make sense. They were explaining less than 1% of the variance, and this can not be the case anywhere in the ocean let alone at 4,000m. After baking my head against the wall for a day or two, I realized that a). the time interval that I was using to run the harmonics on, was not constant, and b). the time that the current meter on the rover was recording was not linear.

Once I figured both of these things out I was able to interpolate, and sort the data into a cogent form.

Interestingly (and spoiler alert here) the various components of the tides still only account for about 20% of the variance.

On top of all the project stuff, MLML had their 50th anniversary shindig on saturday, and I had a lot of fun meeting with my predecessors in the phys oce lab, and hearing about the history of the institution, and finding out what everyone does now.

Last week was kind of a blur. There were a few last pieces I needed to implement and many a bug to squash, but I think I have a working prototype! Carlos, my mentor, set up a wonderful UI to run this language on, incorporating the error checking elements of the parser and validator to provide real-time feedback on errors in the code. Before throwing in the towel on TethySL (the language's name), I do need to update some of the example mission files built in to the UI so they showcase different elements of the language's syntax and how error reporting works. As it is a prototype, there are still some flaws and room for improvement. I will approach these if I have time this week, but for now my focus is almost entirely on the presentation and the report, which I find much more stressful than the actual project.

 Good luck on the presentations everyone!

The pressure is definitely on now that this is (so sadly) our last week. I've been collecting a ton of data to tie all the loose ends of my project, which has yielded both great and not-so-great results. Today is my last day of experiments (you'll see me decked out in black today), so here's to hoping all goes well.

Last week I regenerated some photoproteins from my organisms that I hadn't been able to regenerate previously. I was especially excited when I received signal from a doliolid sample - until I realized that it was quite literally THE ONLY sample that should not have given signal, as it is a species not known to bioluminesce that was collected as a negative control for my other doliolid assays. However, I have determined that the sample clearly has photoprotein in it, which could mean one of two things: contamination (but by what?!) or that the species is actually bioluminescent (less likely). In any case, the truth can't be known until more samples are collected and my experiment is repeated with them.

Aside from that, my radiolarian and phaeodarian data looks good, and today's experiments will mark the close on my research with them (only for now, I hope).

Can't wait to share and hear about all of your projects on Wednesday, this will be fun!

Last week was pretty hectic and stressful, but all in all it was still a good week and I am constantly learning! I spent all of last week working on my presentation and paper. I was a little frustrated because I put a lot of work last week into a section of my research that I will no longer be using in the presentation or paper, so I feel like I wasted a little bit of time but that is ok because I still learned things in the process. I also realized last week how much time and effort go into the editing phase of presentations and data analysis. I spent most of the weekend working on the final touches for my presentation and practicing my talk. I just hope it goes well because I am extremely nervous! I hope everyone has a wonderful final week and that you get everything completed in time!

Last week I continued to process the raw accelerometer signals in an effort to obtain a more accurate velocity estimate from the experimental data. In the past I have obtained great when the translation stage motion was progammmed to have a velocity of 200 mm/sec and a constant accel/decel of 150 mm/sec^2. However, based on the dynamic model of the CPF I created, we believe that the accelerations will be three orders of magnitude smaller. So I have been continuing to refine my signal processing work flow in order to obtain the best possible results (in the remaining time that I have) using the very small acceleration of 1 mm/sec^2 (the lowest that the linear translation stage can move at under the current configuration settings).

Gene also suggested that I test a frequency selective filtering process by creating digital filters with the Matlab Signal Processing Tool Box. The idea was to process the raw accelerometer data with either a high-pass or band-pass filter to remove background noise and the trend before using a cumulative integration method to estimate the velocity profile. In order to systematically run a series of tests and quantify the results of various filters/filter settings, I created a script to plot the average velocity values and variance error bars. In this was it was possible to compare the moving average filter, denoise filter, and new frequency selective filters all together in a single error bar plot. Based on the results from working in the lab on Friday, it seems that the moving average filter and denoise filter offer the best results with the current data set, however with more time I believe that the more general frequency selective filters would be the most robust solution since an optimally tuned band-pass filter could remove the noise and trend from a signal without the need of the detrend() matlab function. For now, I will continue to use the moving average filter and denoise fitler to finish processing my data so that I can complete the presentation and internship report.

Well, I broke my streak of doing Confluence on time, but at least I don't think I'm the last one to post (am I?).

Last week was a crazy race to finish my actual data analysis. I finally got all my data, since I am comparing one dataset that I produced to another dataset that was produced by an MBARI collaborator. Once I had both, my mentor helped me filter them so that they were in the same format (we found the sample number of the max amplitude of the click), so they could be adequately compared. With that data, I did a number of analyses to find out different things about my data.

Originally I decided to see if I could use overlap to count matches, but in addition to perhaps not working just because of the sheer number of samples (256,000 per second, so there's a lot of room for variability with a signal lasting only about 100 samples) I also found out that MATLAB doesn't actually like having 900 million data points in one array to work with. Go figure. I did the whole analysis, I just don't really have a way to look at results at the moment. Luckily I'm going to get to stay on for next week since I missed week 1, so that's something I can look into then.

I also made a filter to get rid of the 50 kHz hum from the MARS power supply, but although the filter worked, it did not produce the results I wanted, because PAMGuard actually did worse on picking up only clicks instead of random noise. I talked to my mentor and Paul McGill about it, but we couldn't figure out what was wrong. In the interest of finishing my project in time for the symposium, that's another thing shelved until next week.

I finally did all my analyses on Thursday and Friday (MBARI is very creepy when no one else is around- there were plenty of people in building A but only like one other in building B so every time that person moved around I jumped) and got weird but interesting results. I've spent the last few days checking and rechecking my data and thinking of new ways to represent the data in a way that's less confusing. As I'm working on my presentation I'm realizing the full extent of how much my intimate familiarity with the data makes my graphs more coherent to me- I'm simplifying, simplifying, simplifying. So hopefully no one will be confused by my data by the time my presentation makes it to its final form.

What a busy week! Sorry about the intense lateness of this update.