Week 4: The Dive Continues

Salutations, intrepid reader! Thank you for returning to my blog as I try to surpass a hopefully rather small bump in my Senior Project journey. This week, I’ve continued my work from last week, attempting to determine the relationship between a curated selection of shark incident features and their potential influence on the inciting of these shark incidents (from the Australian Shark Incident Database), and bettering my understanding of the significance of the number of yearly shark catches (from “A Review of the Biology and Status of White Sharks in Australian Waters”). Additionally, I wrote some emails and messages to researchers and conservation groups with the hopes of interviewing them about their works which have been essential in my project’s development and/or about their contributions and observations to the global front of marine preservation.

A Struggle Against the Current

While last week I focused on Multiple Correspondence Analysis, this week I tried my hand, or data, on Multivariate Analysis (MVA) and Cluster Analysis. As a brief rundown, MVA is a rather broad category of numerous forms of analyses that is mainly intended for use with multiple sets of quantitative data where certain “variables” can be grouped on a variety of axes and then compared to see where each variable or, in this case, shark incident factors overlaps or differs. Cluster Analysis is a specific variation on MVA where we specifically try to group these similarities into separated “clusters” using certain characteristics found in the data points.

Cautiously hopeful, I pushed through various module documentations and forums to figure out how to make my data work, but once again, resurfaced empty handed. The tutorials I found on MVA all, predictably, revolved around quantitative data, and even after converting my categorical data into quantitative 1’s and 0’s using the one-hot encoding method I learned last week, I arrived at first, a graph whose syntax did not fit any of the modules that were said to be required and then, a rather inconclusive heatmap that may foreshadow the final results of my attempts at cracking the Australian Shark Incident Database.

In the later end of this week, I also tried running Cluster Analysis on the database and ended up with a graph that not only I don’t quite understand, but also a graph that looks nothing like what Cluster Analysis usually looks like (search up images and compare them to my graph, and you’ll easily see the difference). Even though this is concerning, I am not giving up on this method yet as there’s definitely some potential (and it didn’t break down on me yet).

In more positive news, I was successful in creating some more insightful versions of the significance graphs of the various catch programs from last week. In particular, I added the 30% error bars to the numbers of sharks caught graphs and created a running average graph, both of which included the color-coding system I created last week (red for “significant increases” in sharks caught and yellow for “significant declines”). These graphs should help to make it visually clear why particular years stand out as outliers from previous year trends.

(If you’d like to take a look at any of the graphs I refer to in this section, here is a Google Slides link to the images: https://docs.google.com/presentation/d/1eH9r4rMJK-ODqEy7nRLR4lgvD0iBH1Ud8CzwBd-T78M/edit?usp=sharing)

Searching for More Opinions

Beyond just messing around with plots, I am planning ahead for future weeks by sending invitations for interviews from researchers. As for the general themes for these interviews, I intend to ask about the researcher’s or organization’s current and past works, sharks and their roles in their ecosystems, shark sensationalization, and shark attacks. These expert opinions should clear up any suspicions I have and better inform any conclusions I hope to make regarding my project. Of course, I have written down nuanced questions for each individual, but I’ll just have to wait and see where my cold calls take me.

Returning to the Depths

Looking ahead into the following week, I am crossing my fingers that something works for connecting the factors I selected from the Australian Shark Incident Database. However, I am also ready to let this part of my analysis go if it proves too time-consuming as the main thing proving any or no connections (which may be what I have now) with features of shark incidents would be offering advice or clearing up any common misconceptions associated with shark attacks. Overall, this proof would be a neat addition to the promotional video I plan to produce, but it would hardly be the end of the world if I can’t draw anything conclusive.

Additionally, I am aiming to find possibly influential, historical Australian written, audio, and/or visual reports, shark (control) policies or programs, and other publications or records that could add some context to my previously noted significant time periods. This upcoming week should be the start of the creation of a sourced timeline to potentially explain how popular portrayals of sharks came to be and why the numbers of sharks caught and/or Catch per Unit Effort (as a reminder, this means the numbers caught per hour, net, etc.) in these shark control programs or fishing groups changed over the years.

Citations

Malcolm, H., Bruce, B. D., & Stevens, J. D. (2001, September). A Review of the Biology and Status of White Sharks in Australian Waters. CSIRO Marine Research, Hobart. https://publications.csiro.au/rpr/download?pid=procite:1d0d13e5-7a60-4e65-be78-636e6f2dd22e&dsid=DS1.

Meagher, P. (2024). Australian Shark Incident Database [Data set]. Zenodo. https://doi.org/10.5281/zenodo.10476905.