Programming Homework Help

Utilize LARGE “R” Dataset of Congressional Documents: https://drive.google.com/file/d/1Q_-EUwW2I_C9WrYCgpxlcIrYfoMTx3os/view?usp=sharing

RESEARCH Question: Does the participation or discussion of UNION by a speaker or state reflect differences geographically or between parties?

PROVE Hypothesis: The substantive affect of geography has a stronger influence than party affiliation on how favorably or unfavorably members of Congress discuss collective labor rights regarding unions.

R Code MUST include the following:

-Load in 2018-2020 Workspace R Data included from Google Drive

-Create a DFM / CORPUS

–Define how many documents are included

-Determine TOKENS and FREQUENCY

-DFM_LOOKUP Function for keywords: UNION, Scab, Strike, Collective Bargaining, TAFT

-Segment the text by Speaker

-Create KWIC Keyword Windows utilizing UNION, Collective Bargaining, TAFT

-Generate a relatively large kwic window around the key word UNION

-Create GLOVE Word Embeddings codes

-Run a REGEX code to view ALL instances of UNION or UNIONS

-Run a Sentiment Analysis through coding

-Use STM to TOPIC MODELl the text in the kwic windows, using a prevalence variable of your choosing

-Use estimateEffect, barplots, and any other methods to illustrate differences in topic proportions across your speeches

-Include UNION Strength statistical information to pair with the DFM: Union Membership from US Bureau of Labor Statistics ( https://www.bls.gov/news.release/union2.nr0.htm ) as well as pertinent data from (http://unionstats.gsu.edu/)

-Run any other code & models that defines the hypothesis, UNION STRENGTH, UNION MEMBERSHIP including histograms and visualizations

After the above R Code DataFrame is initiated, must complete the following:

-Include code and explanatory #comments for the following steps:

• Using either a continuous or binary dependent variable, run the appropriate regression, generate an output table, and interpret the results within the context of your research question.
• Now run regression for predictive purposes. Generate the relevant measure(s) of predictive performance and assess how well your model performed. Experiment with using more/different x’s and observe the difference, if any, in predictive performance.
• Add interpretation and explanation throughout your code with liberal use of #comments.

**R Code & Reporting of Results w/ Visualizations**