I generated the data using hydrodynamical simulations. I used Python to analyze the output of these simulations (HDF5). I made the plots using matplotlib, and the movies by generating each frame in matplotlib then putting them together with ffmpeg.

This data set is comprised of 60 predator-prey or top predator-mesopredator couplings. For each coupling, the temporal activity for each trophic level was recorded in both low human disturbance and high human disturbance conditions. We then analyzed that data and calculated the change in the factor of temporal overlap in R, and plotted the change of overlap for each pairing. This data visualization will be incorporated into a future publication on how human-based activity affects temporal cascades throughout all trophic levels.

I created a website in Visual Studio Code, GitHub, CSS, and Javascript. On the webpage, I uploaded my data visualizations: (1) The graphical abstract shows that the estimated catch rate of by catch from these commissions does not match the data stated from these companies of their bycatch—with only one of the affected species (Silky shark) illustrated and highlighted. I Illustrated the Silky shark in Adobe Illustrator and Photoshop. I also drew the map in photoshop. I created a 3D animation using a free, pre-rigged 3d model of a shark from the site turbo squid (thank you to user Flytet) and the animation software Blender to illustrate how sharks get caught as bycatch.
Third, I used the software R studio and ggplot to take the given CSV data of the number of affected species landed or discarded, and turn it into a graph. I also used ggplot to color code the species based on how endangered they were (using the IUCN scale).

I worked with data from the Scott Creek Life Cycle Monitoring Station which is in form of a CSV. I had to clean the data to only keep the columns I was currently interested in displaying and got rid of rows that had incomplete data. The goal of this project was to develop a Shiny App which is done through R and RStudio. Since it was my first time using R, I had to familiarize myself with the syntax and packages before diving into the development of the app. I use packages such as tidyverse, dplyr, patchwork, and ggplot2 to work with and plot the data. I have user interface and server functions through the shiny package; these control the aesthetic and content of the interactive website. This interactive website will continue to be used and improved by the Landscape and Seascape Ecology Team at NOAA. Displaying the data will allow others to see which fish populations have declined and recovered in the past, identify resilient and sensitive habitats, and more.

The data was shared with me via Google Sheet, where I exported it to a CSV. After creating my HTML page, and linking both my Javascript file and the D3 library, I had D3 parse through the data, and organized data fields from the CSV into a array of Javascript objects that would fit my plans for the data.

I then fed my new data structure into the creation of an Scalable Vector Graphic (SVG) window, where SVG circles were drawn only if they fit a certain description ( e.g. the data was from a certain year/month, the data includes a location). After the program decides which data gets a circle drawn, I used D3's force simulation feature to have the circles be pulled towards a particular point in the window, based off of each data point's location data.

I wouldn't be surprised if resolving bugs and errors comprised of at least half my total dev time. Learning D3 norms, quirks, and syntax in chunks, was a very interesting experience, and internet forums as well as tutorials were immensely helpful. I was already very familiar with web dev (HTML/CSS/Javascript), however, learning a new library was challenging.

We first converted our sea surface temperature data using an NetCDF converter in Python. Before diving into an implementation, we worked with Mer to discuss design in order to maximize reader-friendliness. Within ObservableHQ (JavaScript) we used d3 to wrangle the data, and used the JavaScript Vega-Lite API to render the data as a user-interactive graph. This step was difficult because of the sparse Vega-lite Javascript Documentation! We then used d3 and plain HTML to create accompanying SVG's in order to emphasize important information. At this point, we realized we had readability issues. This led us to completely rehaul the figure's legend; as well as the SVG's. The last step is to leverage colors in order to improve readability!

Stable isotope data were compiled in Excel and converted into CSV files. Individual CSV files were imported into R to create the density plots for each species using the libraries “ggplot2” and “ggridges.” The color of the clusters for each species was manually assigned. The map was generated in R using the libraries “sf” and “ggspatial.” Once generated, the several plots and the map were exported in PDF and then imported into Illustrator, where they were compiled to create the final figure. The otariids icons were vector graphics created in illustrator and then colored and converted in raster graphics in Photoshop to then incorporate them in the final file in illustrator. This data visualization will be used for the publication of one of my dissertation chapters.

I ran simulations using R on the supercomputer hummingbird. I then would then take the output and process the data in R. I made both the website and all my figures in R.

To make this video, I used PowerPoint to animate raster graphics depicting headwater streams and invertebrates. When designing the video, I employed graphic design concepts from the Data Visualization course. This video will potentially be used in a future museum display about freshwater ecology.

Adobe Illustrator was used to create a graphical abstract. The abstract consist of two diagrams that show the different juvenile migratory strategies observed among the successful adult spawners under wet and dry conditions, and the additional effect of water temperature increase on the amount of cold water pools available for juvenile salmon rearing.

Data from time-depth recorder and satellite instruments were processed in MATLAB and converted into KML files. Individual KML files were imported into a single KMZ zipped file to be imported into Google Maps' "My Maps" tool, which allows users to quickly and flexibly build interactive map interfaces. Metadata was exported from MATLAB into a CSV which was then paired to the data layer containing each respective animal's track. This interactive data visualization will be incorporated in an Op-Ed we are writing for the Conversation to accompany our latest publication in Science Advances "Lightscapes of Fear: How mesopredators balance starvation and predation in the open ocean"