Mark Yamane

About Me

Hi, I'm Mark! I am an Eckerd College graduate from Shoreline, Washington with Bachelor of Science degrees in Computer Science and Marine Science. My interest in the ocean has driven my pursuits in both ocean research and data science, so in my free time you can find me hiking around Washington, scuba diving off of Edmonds, or taking low-tide beach walks in West Seattle. I made this site to showcase some of my projects and past research, so feel free to check them out below!

Projects

Lionfish

biod

I created an R package out of functions I used to perform biodiversity analyses on data from visual dive surveys I conducted with a group for my Winter Term in Roatan, Honduras. We conducted a study on the effects of lionfish derbies on the native predator fish assemblages (which now compete with the invasive lionfish). This package has functions to find species richness, species evenness, and Shannon diversity index for uni-, bi-, and multivariate data from a csv file. 'biod' also includes some functions for graphing the diversity values.

Tech Used: R (dplyr, ggplot), RStudio

Code: Not Available

Demo of Raxin

Raxin

I created a discord bot called "Raxin" to facilitate reaction polls in Discord. I was the president of Tabletop Club at Eckerd College when COVID hit the US in March 2021, so when Tabletop Club had to go virtual I chose Discord as the solution. When we split out into groups or were deciding what games to play, it was often way too chaotic to go off of the chat or audio conversations. I decided to create a bot to easily and intuitively create polls where people could respond using the given reactions attached to the polls. This bot was used many times when we were virtual and still when we phased back to in-person meetings.

Tech Used: Javascript, Node.JS

Code: https://github.com/mtyamane/Raxin

Picture of this website

This Website

I was assigned to take "Creation of a Personal Web Presence" for Autumn Term my first year at Eckerd. This class taught how to create a website using HTML and CSS, as well as incorporated discussions about ethical problems in the social media sphere. During sophomore year, I decided to take some time to learn more about CSS and JavaScript by making this site!

Tech Used: HTML, CSS, JavaScript

Code: https://github.com/mtyamane/mtyamane.github.io

Research

Tortoise results

Tortoise Recognition

Re-identification is a crucial time-consuming task for many studies on behavioral ecology. For my senior thesis at Eckerd College, I developed a neural network with Dr. Michael Hilton to identify individual gopher tortoises based on carapace markings in a manner similar to how facial recognition is used on humans. The model calculates a difference value to return the five most similar gopher tortoises from a reference database. Once the model achieved 95% accuracy, we incorporated it into existing camera trap analysis software to aid future studies defining the social networks of gopher tortoises at Boyd Hill Nature Preserve.

Tech Used: Python (TensorFlow, Keras, NumPy, Matplotlib)

Publication: https://arxiv.org/abs/2206.05159

Code: https://github.com/hiltonml/camera_trap_tools

Picture of saildrones

Arctic Weather Model Validation

As sea ice is diminishing at unprecedented rates, accurate weather forecasting of the Arctic Ocean is pertinent for management and conservation of marine resources and coastal communities. During an REU at the Pacific Marine Environmental Laboratory with Dr. Chidong Zhang, I validated ensemble forecasts from the European Centre for Medium-Range Weather Forecasts (ECMWF) against observations from saildrones - remotely controlled uncrewed surface vehicles that measured atmospheric and surface ocean variables. Forecast errors in dynamical variables began small and grew over time while forecast errors in thermodynamical variables began large and stayed large up to a lead time of 15 days. This study provided additional insights to the degree to which ensemble forecasts may be advantageous to deterministic forecasts within the range of 15 days.

Tech Used: Python (Xarray, NumPy, SciPy, Matplotlib)

Publication: https://agu.confex.com/agu/fm21/meetingapp.cgi/Paper/961273

Code: Not Available

Demo of Raxin

Harvest Cessation in Marine Reserves

Determining whether illegal harvest (poaching) has continued in a protected area is important to planning enforcement and adaptive management. During an REU at Hatfield Marine Science Center with Dr. Will White, I used a state-space integral projection model to estimate harvest rates for four kelp forest fish species inside marine protected areas (MPAs) and non‐MPA reference sites in the California Channel Islands, from 2003 (when MPAs were implemented) to 2017. This modeling approach could provide a tool to complement the long‐term management of MPA networks, particularly given the difficulty of acquiring harvest rate data at the spatial scale of individual MPAs. This internship resulted in a co-authored paper published to Conservation Letters.

Tech Used: MATLAB, R

Publication: https://conbio.onlinelibrary.wiley.com/doi/full/10.1111/conl.12775

Code: https://github.com/jwilsonwhite/IPM_statespace/tree/master/Channel_Islands