claire-xuey1 / big-o-of-analytics

Analysis Project-2018

Big-O-of-Analytics

Analysis Project-2018

Deadline: Dec 4th

Background:

The ACME corporation is considering moving into agricultural inputs in Ghana. To get an idea of whom to target for their sales efforts they have hired you to analyze what determines agricultural profit there. They are especially interested in what effect household educational attainment and the characteristics of the local area has on profit, but if you find other results of interest, they would like to hear those as well. They have left the choice of explanatory variables and variable definitions to you, but to be able to compare across regions they want at least one specification that examines profit per area unit (for example, acre, hectare, or square meters)

Data:

ACME has provided you with access to the Ghana Living Standard Survey 4 from 1998/99, including all the documentation that comes with the survey. The data files are in Stata format and contain the raw data from Ghana Statistical Service, so expect a fair number of data issues. Not all the data are equally relevant for your work.

You can find a zip file with all the data and the associated documentation here. Make sure you read the documentation before even looking at the data. Note that some of the data are calculated aggregates from a report put out by the World Bank. You can use these as well. Documentation for the aggregates is also provided.

Memorandum:

Write a brief (500 words or less) summary of your findings in standard memorandum format based on your analysis. ACME's board is not interested in the statistical analysis; they want to know the results and your interpretation of them so that they can use them when deciding on their plans. You must interpret and write up the results in a form that makes sense to them; do not expect them to have much statistical training. The memorandum should NOT contain technical jargon. You must include clearly labeled and explained table(s), graph(s), or both to highlight relevant points about the data and statistical analysis.

Statistical Analysis Paper:

The statistical analysis paper is for internal discussions of plans, and it is possible that ACME revisits your analysis. It is therefore imperative that your statistical analysis is clearly and thoroughly documented. You must describe each step you take and why. Methods, hypotheses, and results of the analyses you do should be clearly explained and labeled in the appendix.

R code:

Furthermore, you must submit an R project with all the files needed to reproduce your analysis (except for the original data). You can use GitHub, or you can zip the entire project and upload it. It is essential that your project can run on another computer (as long as the end user installs the necessary packages and the relevant data). Remember to document the code thoroughly. Do not change the names of the raw data files.

Grading:

Your grade will depend on your ability to:

identify relevant variables to address the overall question recode and create variables as needed/appropriate deal with untidy datasets combine data sets perform model assessment/diagnostics as appropriate present results in a technically sound manner present results in a way accessible to a non-technical audience assess other groups' work