diff --git a/ExpectationsStudents/expectations_cl.md b/ExpectationsStudents/expectations_cl.md new file mode 100644 index 0000000000000000000000000000000000000000..c6f6e698db4c76734de1de6643814a5349159a14 --- /dev/null +++ b/ExpectationsStudents/expectations_cl.md @@ -0,0 +1,40 @@ +# First Assignment + +## Class Expectations + + +I assume this just is supposed to be some kind of lorem ipsum placeholder text. +However, in case it is not here are a few words on my expectations of this +class: After finishing my Master's thesis in political science last summer, +I felt that the approach I pursued within this thesis, which can be characterized as +computer-driven text analysis which was used to determine political preferences of political +actors on the basis of speeches they delivered and press statements they +released, had some potential if applied the right way. Packages for r like +*wordfish* enabled me to do so on my own. However as an autodidact (and way more +an political theorist than a computer scientist) both the preparation of the +text corpora and the analysis had some flaws which made things like text mining, +modelling as well as the interpretation and modification of the resulting tables +and graphs more cumbersome than probably necessary. Coming from this starting +point, I hope this course will make it possible for me to use the things I +already know in a more structured, informed and faster way. Of +particular interest for me are data visualization principles. After producing +long essays and theses for a while now, I certainly appreciate a well made graphical +representation of complex data. + + +### Autodidactic Basis + +Websites I use as a self-taught r-user interested in text analysis: + +* quantitative text mining on [textasdata.com](http://www.textasdata.com/) + +* Some fancy applications of r, both descriptive statistics and more advanced + stuff on [r-bloggers](https://www.r-bloggers.com/) + +## Top Expectations + +* No more hand-cleaning of 2000 parliamentary speeches + +* Being able to have a look under the surface of r-packages. How does topical clustering work from a computer-scientist's view and how do I avoid weird outcomes and bad graphical representation? + +* Better Visualization! Maps in r! What does aes do?