What a week! It has been a lot of fun talking to the TCLers at MBA and really good practice for educational speaking. I haven't gotten a lot of time to work on my research so I know that this evening, tomorrow and Sunday will be pretty crazy in regards to working on my MBARI project. I have begun working on my paper and I am still waiting for the final pulse images to be uploaded to VARS so I can finalize my data collection.

I had some pretty amazing conversations this week with MBARI staff, Tina Swanson, MBA staff and the other interns. I think I am getting a more solidified idea of what steps I would like to take in the future regarding my career and I feel very fortunate that I have been able to develop these ideas during this internship. I do wish there weren't so many interesting things to do in life, there just isn't enough time to do, see and learn about things!

Week 6:

Project update: I haven't yet looked at my stained samples from the cruise on the nice land microscope, so that's still to be done. I have been working on 3D reconstruction in MatLab a fair amount, but if anyone can help me understand linear algebra, that would be a lifesaver! That's the biggest roadblock right now to getting a 3D model from the SEM images. I found example code that I think I can modify for my purposes, but it's very old and uses functions that don't exist anymore, which is a challenge to figure out on top of the math itself. I want to focus my project on the 3D modeling and take a closer look at some of the smaller rads from the cruise as initial examples...much to do!

Fortunately/unfortunately, I haven't had much time at my desk this week! Monday, we returned from the cruise, then Tuesday we spent the morning offloading lab equipment and went on a behind the scenes tour of the aquarium in the afternoon. Wednesday we spent all morning preparing food for the barbecue (a great success! Go Drew!), and today I'm presenting to the TCL group at the aquarium! Maybe I can make up some time on the weekend, or just read up a bit on linear algebra to make work go faster? We shall see...

On time, Oh Yeah!

The BBQ dominated my week, and as such I've not gotten as much done on my project as I would have liked, but c'est la vie.

I have finish writing my machine learning algorithms in Matlab, and they seem to work pretty well with both the made up data sets that I give it, as well as the actual rover data, which is really encouraging. I know have to translate the code into ruby so that the rover will be able to utilize it.

I'd still like to build off the machine learning code and work on some fault detection with the rover, but the internship is almost over and it may to much to attempt in such a short time.

In addition while I was out on the western flyer with the smith gouger I was able to play around with some of the rover current data and discovered a few things that I think will be very valuable to the team going forward. I would also like to summarize these findings regarding the physical environment and present them to the team along with some recommendations. A lot to do in not so much time.

Week Six:

Wow, the weeks are just flying by. I have had a few discussions with Ben and Brett this week about project Ideas. At the end of last week Ben and myself met with Hans Thomas from the Dorado group and we discussed potentially applying the Xg measurement and moment arm calculation procedure to Dorado, we were able to have Hans agree to partially fund the scales we need for the project. I spent some time reworking my written procedure and am still working on a more general AUV procedure as the LRAUV has certain advantages which simplify the process. I have been working on chock design ideas as well as the frame assembly for the scale positions to try and decide on the best designs. With the end of the internship quickly closing in, I hope to have the scales arrive and my LRAUV chocks, and scale frame all complete by the end of next week. I still need to work on drafting up a spreadsheet which takes inputs in a sequence and calculates the desired values. Just hoping it will all comes together in a timely manner.

So the data that I'm working on is all cleaned up and null byte free! The next thing to do is make a loop to have the program read through all of the float numbers and, via thredds_crawler, read all of the profiles for the floats. So this is where I start getting into the data reading stuff where I will finally be able to do some calculations! Sweet B) then from there I will read the World Ocean Atlas data where I will compare it to the Argo data in order to calculate a correction factor. This will then be applied to all of the Argo data to have a quality controlled data set, something the oxygen data lacked and is the reason why I'm working on this!

Hello all you beautiful people.

This week has felt as though it's been mostly dominated by deadlines for Hartnell. I've had to send in my project abstract as well as finish the rough draft of my poster. The poster is looking nice so far (at least in my opinion). For my project I'm finding myself installing more applications for Django and STOQS. Of course, these installations are fussy and need additional commands in order to work.
For the fun events this week! The MBA tour was very interesting, but it would've been better had we all been given complimentary otter pups. It's okay though, I'm only a little heart broken. The BBQ was delicious (all thanks to Drew). Also, shoutout to Paris for his charitable work in the cleanup crew. Not all heroes wear capes.

Week 6:

On Monday, I talked to Jen who is in charge of the data sources that CeNCOOS links to from their website, and we talked about two datasets from NOAA ERDDAP that could easily be brought in to the CeNCOOS ERDDAP and linked from the CeNCOOS site, CalCOFI and the NOAA rockfish surveys.  (I did look through the Rockfish data on Week 5 and realized it only has 4 years of data and I won't be including it in my data (at least not yet) because it will not add a lot of new insight).  On Tuesday, I finally finished a regression plot of oxygen over space and depth.  It tells me where along the Line67 transect the decrease in oxygen over a 15 year period is the greatest.  I also talked to Reiko of the Chavez group about outliers and threw out a bunch of outliers in the data I'm working with, as well as provided Reiko the list of suspect values (and their cruise, ect).  On Wednesday, I worked on a regression plot over potential densities over space.  I had to throw out more data from my analysis.  On Thursday, I finished the regression plot over potential density.  On Friday, I will likely work on the ROV data comparison, which is a leftover task from Week 5.  I also may to model output analysis along Line67.  If not, I will do the model analysis this weekend from the dorms.

I have also begun learning Python this week (thanks Prateek).  We have spent some time working at the dorms (again, thanks Prateek!).  I also drew myself a phytoplankton key the previous weekend in preparation for Open House - and rewatched some Plankton Chronicles videos from the TARA expedition.  Hopefully all of those phytoplankton genus names will stay in my head until 5 o'clock on Saturday.

Haha nice look on the shoutout Olivia.

Yesterday I went on my Rachel Carson cruise and spent the morning helping the crew clean out the refrigerator and pantry. I also did a good amount of sleeping and watching "How I Met Your Mother"... awesome trip!

Can't believe I planned a month ahead and still got stuck doing kitchen duties.
Cleaning up is lame.

My poster is just about finished.
Stop by the Geology tent during Open House to see it!!!

I completed the 4-20mA current loop transmitter test, the thermal test and the parallel test. The instrument simulator can now communicated to the terminal with 2 ports. I am able to simulate all ports at the same time. The project is more complete now. A lot of problems were solved through initializing the setup.I am going to start on my paper and research on ways to improve the design and find theoretical ways to support my design.

I got data back from sensor number one this week! The data looks good and I am currently waiting for dissolved organic carbon data to be able to validate my data. I am still in the process of building the second sensor, which will hopefully be done sometime before the next confluence. In unrelated news, I now have a smaller brace, which makes typing and soldering much easier! Last, my mentor is gone next week to go to NYC for the X prize announcement.

 Had many brainstorming sessions with my mentor this week, so just love the progress in my project. We have finally decided to work on two different classifiers working in parallel to solve the same classification problem. The only difference is one classifier is an acoustic signal based classifier while the classifier that works in parallel is a deep learning image based classifier which works on the spectrogram images of the acoustic signals(calls of the different species). We hope to get better results and performance with this kind of architecture. Also having classified blue whales and fin whales , I am introducing more diversity into my classifier( right now Cuviar's beaked whale and porpoise). Also this week I got a chance to meet the people at Hopkins who study whales using acoustic tags by tagging them and studying the data from the tag.

I am late again with my confluence and I feel so bad! I apologize. My time here is going by incredibly fast, sometimes I forget what day of the week it is. Time please slow down!

However, this week I have been able to make progress on translating Brian's scala example code into Matlab. It is all ready to run, all I need are the camera parameters, which I have two out of four. The other two: alpha, the height of the image in radians and beta, the width of the image in radians require a bit of math to solve. I currently cannot finish that math because I have to get the distance of the ROV camera from the calibration board when it was on the Western Flyer. Hopefully, I will be able to get that measured on Monday when the Doc Ricketts is back aboard the ship....and hopefully grab a sweet Doc Ricketts t-shirt while I am on board

Week six:

In this week I could finish to watch all my videos. I joined to another stage of my project where I needed to get all my data, clean and start playing with it. I am finding interesting things that are happening in the deep sea in the Gulf of California.

On the other hand, the MBA tour was great. Have the opportunity to see everything behind the scenes is a privilege. Also, the BBQ was excellent. It was a mixture of tastes, smells and colors. Even Cuban music was the best complement to this activity. Thank you Drew for your wonderful idea.

I'm at the scope counting phytoplankton slides we collected this spring. The eDNA samples were shipped to Stanford for sequencING and we await the results. The intern bbq was a success. Open house also went really well. Quite a few people were asking some great questions about the plankton. And thanks to Brian for hosting a Matlab session.

Last week was pretty busy as I began processing all my gas spectra. On Monday, Tuesday, and Wednesday, I spent a lot of time searching the literature to identify unknown peaks in my sample spectra, which appeared mostly to be different vibrational modes of hydrocarbons I already knew that I had, plus CO2. On Wednesday, I really enjoyed meeting Tina Swanson. I have been considering pursuing advocacy later in my scientific career and have read a good amount on NRDC, so she was really great to talk to. I've been through all my spectra with one version of my Matlab code to integrate all the peaks (manually setting integration limits each time) and started integrating everything with another version of code. I'll have to see which version gives more precise results.

Last week was a very busy week. I got to go on the Rachel Carson which was amazing! I am still trying to refine the code to pick out more jellyfish accurately in the videos. Right now the code is picking up things that are not jellyfish so I am in the process of writing the code to make sure it doesn't count anything as jellies with areas smaller than a certain number. Once I get this I can really compare the manual and code data together to see the accuracy!

Alaska was amazing but it feels good to be back in the swing of things and on a routine. This week has been marked by a very crucial decision to terminate automatic kinematic extraction and instead turning to manual extraction with a bamboo pad. This decision was made due to the finicky nature of the videos from past ROV cruises and Matlab's unwillingness to compromise. While I am a bit disheartened, I am also excited because I know that manual extraction will be successful which means one step closer to results.