usf-portal · bpearson11 · Nov 13, 2018 · Nov 13, 2018 · Nov 13, 2018 · Nov 24, 2018
diff --git a/README.md b/README.md
@@ -3,38 +3,16 @@ Digital Workbook, [University of South Florida](http://www.usf.edu/), Fall 2016
 
 ---
 
-Instructor: [David J. Thomas](mailto::davidjthomas@usf.edu), [thePortus.com](http://thePortus.com/)
+By: [Brandon Pearson](mailto:bpearson1@mail.usf.edu)
 
 ---
 
-<figure>
 
-![Replace Me, Sample Image](docs/imgs/caesarian_code.png)
-
-<figcaption>*Put a caption to your image here, if you want*</figcaption>
-
-</figure>
 
 ---
 
 ## Project Template
 
-This is a starter template for final projects. When you have completed your final project, you should replace this message (README.md) with a short 1-2 paragraph description of your project.
-
-See the [Course Workbook Project Page](https://hacking-history.readthedocs.io/project) for more information on the final project.
-
-**REQUIREMENTS BEFORE STARTING**
-+ [GitHub account](https://github.com) created
-+ [GitHub Desktop](https://desktop.github.com) client installed
-+ [Atom](https://atom.io) editor installed
-+ [Python 3.x](https://www.python.org/) installed
-
----
-
-## Past Project Examples
-
-* [Confederate Memorials](http://confederate-memorials-project.readthedocs.io/)
-* [The Slave Ledges](http://slave-ledger.readthedocs.io/en/latest/)
-* [An Teanga Sean: The Celtic Languages](http://an-teanga-sean-the-celtic-languages.readthedocs.io/)
+This project consists of my analysis of Hillary Clinton's emails using Orange3's hierarchical clustering widget to group similar emails. To allow the hierarchical clustering to work, I first had to use Orange's "Bag of Words" widget to give the emails numerical value - in this case, how often specific words appeared - and put it into the "Distances" widget to see how similar certain emails were. After running it through those widgets, I chose to analyze cluster C1 - a set of emails about Benghazi - to see what linked them together, finding that the only reason Orange grouped them together was the fact they all contained references to the attack in Benghazi or Libya. Some were about Libyan concerns regarding how it would affect international investment while others were about the 2012 Presidential election; specifically, Republicans' desire to use Benghazi to paint Obama as weak on foreign policy.
 
 ---
diff --git a/docs/background.md b/docs/background.md
@@ -1,25 +1,19 @@
 # Project Background
 
 ---
-
-Put a general discussion of your topic here. Don't get into the historiography or scholarship here. Just introduce to your reader what your topic is. What is the 'problem', the 'question', and/or the 'argument'.
-
-YOU SHOULD BE LIBERAL AND INCLUDE ANY AND ALL LINKS WHEN RELEVANT
+ I chose to analyze WikiLeaks' archive of Hillary Clinton's emails to try and learn about Benghazi, and chose to analyze her emails because it is a political topic, and I am interested in politics. What I learned is Libya's president at least initially believed militants had infiltrated protests against an anti-Islam video to attack Benghazi; I also learned of his concern foreign businesses would see the country as unstable if he was unable to eliminate the militants [(WikiLeaks, 2016)](https://wikileaks.org/clinton-emails/). I also learned of the GOP's eagerness to use Benghazi for their political gain, and the annoyance of Clinton's allies people refused to move on [(WikiLeaks, 2016)](https://wikileaks.org/clinton-emails/). 
 
 ---
 
-## Historical Background
+## Historical and Digital Background
 
-Put a discussion of any relevant historigraphy you think relates to the topic.You can discuss the historiography of specific times and places, but you can also discuss any theoretical literature you think is relevant.
+Part of the historiography of text analysis is the "Gender, Race, and Nationality in Black Drama, 1950-2006: Mining Differences in Language Use in Authors and their Characters" [(Argamon et al., 2009)](http://digitalhumanities.org:8081/dhq/vol/3/2/000043/000043.html). The authors discuss their use of text analysis to test whether techniques can tell the difference between authors of different gender, race, and nationality (Argamon et al., 2009). Though the scholars were successful in their efforts, it was only after normalizing the data that they were able to use the techniques to determine author and character gender, which they fear puts distance between the text the and reader and limit the ability of readers interpret text with text analysis (Argamon et al., 2009). Argamon and his team point out focusing on only simple categories may prevent analysis of complex themes using text analysis or reinforce the biases of authors by "confirming" what they set out to prove (Argamon et al., 2009).
 
-You have freedom to roam from the narrow topic of your project to explore how different authors/schools of thought have impacted scholarly approaches over time. However, make sure that in the end you clearly relate how this discussion relates to the subject of your project and/or your choices in methods or interpreative models.
+Another part of the historiography of text analysis and a rebutal to "Gender" is Dr. Meehan's "Text Minding"  [(Meehan, 2009)](http://digitalhumanities.org:8081/dhq/vol/3/2/000045/000045.html). Dr. Meehan agrees with Argamon and his coauthor' argument technology may force sources to fit models that are too simple; however, unlike Argamon, Meehan states technology could aid research (Meehan, 2009). If scholars familiarize themselves with technology and combine it with traditional tools, it could enable them to find its shortcomings and learn how to circumvent epistemological impediments posed by technology, just as Fredrick Douglas was able to overcome the binarity of slavery by familiarizing himself with and exploring how slavery advocates contrasted slaves with the free (Meehan, 2009). 
 
-As you discuss different authors, you may site them using (author, page) style parenthetical notation. Make sure that a full citation in [Chicago](http://chicagomanualofstyle.org) is added to the 'docs/credits.md' page.
+The relevance of these articles for using text analysis on WikiLeaks' archive of Clintons' emails is that though scholars should be cautious about using techological tools - because they are as flawed as the people who made them  - the way for scholars to prevent themselves from being constrained is to educate themselves on how technology works and combine it with other analytical aids (Aramon et al., 2009 and Meehan, 2009). As I and others familiarize ourselves with Orange3, we will begin to understand the shorcomings of text analysis and digital techniques in general (Meehan, 2009). A better understanding of technology, in turn, will allow scholars to find novel ways to analyze text, and possibly increase the breadth and depth of humanistic study (Meehan, 2009). Therefore, projects using text analysis to study Clinton's emails are useful as they give scholars the ability to conduct more sophisticated research on a subject _and_ the opportunity to familiarize themselves with technology. 
 
----
 
-## Digital Background
+---
 
-You should also make sure to do research on any relevant digital work, whether scholarly articles or digital projects. Make sure to check Digital Humanities Quarterly, or [DHQ](http://www.digitalhumanities.org/dhq/), [Debates in the Digital Humanities](http://dhdebates.gc.cuny.edu/), [JStor](https://jstor.org), blogs and more to find relevant work.
 
-You don't actually have to have a sepearte 'Digital Background' section. If it feels more natural to you, you may combine them into a single discussion. This is especially a good idea if you feel that with your topic you cannot talk about historiography without also talking about digital scholarship, and vice versa.
diff --git a/docs/biblio_and_credits.md b/docs/biblio_and_credits.md
@@ -4,17 +4,32 @@
 
 ## Project Bibliography
 
-* List your bibliography items here, whether historical and/or digital in nature, here.
-* It should be in [Chicago Manual of Style](chicagomanualofstyle.org)
-* Remember to *italicize* and **bold** as appropriate.
+Argamon, Shlomo, Charles Cooney, Russell Horton, Mark Olsen, Sterling Stein, and Robert Voyer. “Gender, Race, and Nationality in Black Drama, 1950-2006: Mining Differences in Language Use in Authors and their Characters.” _Digital Humanities Quarterly_ 3, no. 2 (2009). Accessed December 1, 2018. http://digitalhumanities.org:8081/dhq/vol/3/2/000043/000043.html. 
+
+Brennan, Margaret and Paula Reid. "State Dept. to comply with court order on Hillary Clinton's emails." _CBS News_, May 19, 2015. https://www.cbsnews.com/news/state-dept-to-comply-with-court-order-on-hillary-clintons-emails/.
+
+Currier, Cora and Micah Lee. “In Leaked Chats, WikiLeaks Discusses Preference for GOP Over Clinton, Russia, Trolling, and Feminists They Don’t Like.” _The Intercept_, February 14, 2018. https://theintercept.com/2018/02/14/julian-assange-wikileaks-election-clinton-trump/.
+
+Curk, Tomaz, Janez Demsar, Ales Erjavec, Crt Gorup, Tomaz Hocevar, Mitar Milutinovic, Martin Mozina, et al. "Orange: Data Mining Toolbox in Python." _Journal of Machine Learning Research_ 14(Aug), (2013): 2349-2353. 
+
+Meehan, Sean R. “Text Minding: “A Response to Gender, Race, and Nationality in Black Drama, 1850-2000: Mining Differences in Language Use in Authors and their Characters” .” _Digital Humanities Quarterly_ 3, no. 2 (2009). Accessed December 1, 2018. http://digitalhumanities.org:8081/dhq/vol/3/2/000045/000045.html. 
+
+Orange. "Hierarchical (hierarchical)." Orange Data Mining Library. Accessed December 3, 2018. https://docs.orange.biolab.si/3/data-mining-library/reference/clustering.hierarchical.html.
+
+Thomas, David. “The Portus.” Github. Accessed December 9, 2018. https://github.com/thePortus.
+
+WikiLeaks. “About.” WikiLeaks. May 7, 2011. https://wikileaks.org/About.html.
+
+WikiLeaks. “Hillary Clinton Email Archive.” WikiLeaks. February 2, 2018. https://wikileaks.org/clinton-emails/.
+
+WikiLeaks. “What is WikiLeaks.” WikiLeaks. November 3, 2015. https://wikileaks.org/What-is-WikiLeaks.html.
 
 ---
 
 ## Project Credits
 
-* Put your group member's credits here
-* Link to any emails or github accounts (if you want)
-* Leave the credits at the bottom
+* This project created by Brandon Pearson.
+
 
 ---
 

diff --git a/docs/imgs/workflow.png b/docs/imgs/workflow.png
diff --git a/docs/index.md b/docs/index.md
@@ -1,27 +1,8 @@
-# Replace This Title!
+# An Analysis of Clinton's Emails!
 
 ---
 
-<figure>
-
-![Replace Me, Sample Image](imgs/caesarian_code.png)
-
-<figcaption>
-
-*Put a caption to your image here, if you want*
-
-</figcaption>
-
-</figure>
-
-1. Put an intro image above (if you want)
-2. Change the 1st line of this file to the name of your project
-3. Replace this list with the names of your group members, linking to email or github accounts (if you want)
-4. Remember to also add your credits, introductions/summarys to the mkdocs.yml and README.md and docs/credits.md files.
-
----
-
-Replace this, putting the introduction/summary of your project here. Leave the credits in the bottom section, however.
+This project consists of my analysis of Hillary Clinton's emails using Orange3's hierarchical clustering widget to group similar emails. To allow the hierarchical clustering to work, I first had to use Orange's "Bag of Words" widget to give the emails numerical value - in this case, how often specific words appeared - and put it into the "Distances" widget to see how similar certain emails were. After running it through those widgets, I  chose to analyze cluster C1 - a set of emails about Benghazi - to see what linked them together, finding that the only reason Orange grouped them together was the fact they all contained references to the attack in Benghazi or Libya. Some were about Libyan concerns regarding how it would affect international investment while others were about the 2012 Presidential election; specifically, Republicans' desire to use Benghazi to paint Obama as weak on foreign policy. 
 
 ---
 

diff --git a/docs/initial_findings.md b/docs/initial_findings.md
@@ -1,25 +1,7 @@
 # Initial Findings
 
-This is where you where you discuss any (initial) conclusions you have come to. This is also a natural place to put any and all visualizations that you come up with. In fact, this is where you tell the reader what it is you discovered. You can use the first person if that helps you discuss the process. This is where you both describe to the reader what they are looking at (if they are looking at a visualization) and, most importantly, what your interpretation of the results is.
+My initial finding after analyzing cluster C1 is the only thing tying the emails in the cluster together was the fact each referenced Benghazi, but in the subclusters, there were several common themes, including foreign policy, national security, and the U.S. 2012 presidential election. My analysis of one subcluster revealed to me Libyan President Mohammed Ussef el Magariaf's concern foreign investors would see the country as unstable if he was unable to find the militants who murdered Ambassador Chris Stephens in Benghazi, and his belief several of them were being funded by al Qaedaa [(WikiLeaks, 2016)](https://wikileaks.org/clinton-emails/). My examination of a second subcluster revealed Hillary's inner circle was concerned about the effect Benghazi would have on U.S. policy in Libya, especially projects America had already begun implementing in the country[(WikiLeaks, 2016)](https://wikileaks.org/clinton-emails/). My analysis of a third revealed to me how eager the GOP was to make it an issue (one Romney advisor called it a gift to the campaign) and the frustration of Clinton's supporters the issue did not disappear, one of whom asked why it had not be consigned to the "sad and sealed file of Americans killed abroad in dangerous line of duty" [(WikiLeaks, 2016)](https://wikileaks.org/clinton-emails/). Given I learned all this information from a cluster of roughly 20 emails out of 30,000 total, I believe there is much more that could be learned if scholars chose to analyze the data further; additional analysis of related clusters might reveal some of Benghazi’s effects on Libya's economy and relationship with America. 
 
-If you want to add an image with a caption you can do it like this....
+However, conducting more thorough analysis demands a larger team - this is not a project one person can tackle - with technical skill, and if possible, more powerful technology. Since I conducted the research on my own and due to time constraints, I was only able to look at a single cluster; a larger team with more resources, in contrast, could do a more in-depth analysis of Clinton's emails. Scholars will also need technical expertise as my limited knowledge and skills hampered my ability to analyze the data by preventing me from using hierarchical clustering programs that would have aided my project. Finally, scholars will need better technology - I had to limit the size of the file I ran on Orange so my computer could process it; scholars with more powerful technology than what I have access to would be able to do higher quality research by analyzing a larger sample. 
 
-<figure>
 
-![Replace Me, Sample Image](imgs/caesarian_code.png)
-
-<figcaption>
-
-*Put a caption to your image here, if you want*
-
-</figcaption>
-
-</figure>
-
-Whether you turn your visualizations into static pictures and put them here or embed them, you MUST discuss your visualizations adequately. That means that whoever is the visualization expert must explain what they think the visualization means. You should explain anything that is not self-apparent from the picture alone. Moreover, you should at least comment on whether you think you can draw broader conclusions from any of the visualizations, either when considered individually or all together.
-
-If your project was less visualization-centric, this is where you at least explain in plain words what you learned through the process of non-visual analysis. For example, if you used text mining not for visualization purposes, but to help you manually find interesting threads of conversation in a body of documents.
-
-You should also make sure to be a cautious scholar, and to think about the limitations of what the visualizations can actually tell you. As indicated by the section name, you are **not** expected to make hard conclusions that upend serious historical debates. Rather, this is a place to explore what might be learned from visual exploration.
-
-You must also comment on where you would go next. If these are initial findings, where do you think the best profit would be for any future attempts. This is also where you can talk about pitfalls that limited your ability to learn more about the topic. How might future projects overcome these difficulties?