TLDR; Make something similar to what you did for Capstone 01, but do it with R. Think about something related to inequality, but you can draw from Unit 01 topics if you prefer.
Important: You may work in groups of two to three on this Capstone. If you do, please notify me by e-mail in advance before the showcase so I can adjust small groups as needed.
Capstone 02 Research questions1. Place-based inequality QuestionDataUseful R Commands: 2. Global inequality: QuestionData3. Inequality in labor market outcomes in the U.S.QuestionsData4. OtherSpecificationsDeliverablesOther tips
Select a U.S. county of interest to you - this could be a place where you grew up, a place where you've spent substantial time, or a place that fascinates you (Youngstown, Ohio, anyone?)
Examine the patterns for a number of different groups (e.g., lowest income children, high income children) and outcomes (e.g., earnings in adulthood, incarceration rates). You can do this exploration with the data you have, or in the Opportunity Atlas itself. Then, select a few items to focus on
and/or
Compare outcomes for those from your selected area to the entire United States
Beginner/Intermediate: Use this data set: atlas.dta. Note that you will need to install the haven() package to use the code read_dta() to import it. Restrict your entire analysis to the county you have chosen to focus on - so you will have one observation per census tract for that county.
You will find this data description extremely helpful.
Advanced: You can download either County- or Census-tract data here. You would want to find the data corresponding to "The Opportunity Atlas: Mapping the Childhood Roots of Social Mobility" and then select the appropriate data depending on your level of analysis:
Census Tract: "All Outcomes by Census Tract, Race, Gender and Parental Income Percentile "
County: "All Outcomes by County, Race, Gender and Parental Income Percentile"
You'll also see neighborhood/county characteristics that you can download and merge in.
In either case, you will find it useful to know your FIPS code. The easiest way to do this is within the mapper at Opportunity Atlas. For example, searching for “Lynwood Road, Verona, New Jersey” will display Tract 34013021000, Verona, NJ. The first two digits refer to the state code, the next three digits refer to the county code, and the last 6 digits refer to the tract code.
Calculating unweighted summary statistics
# summary stats, unweighted summary(atlas$yvar) mean(atlas$yvar, na.rm=TRUE)sd(atlas$yvar, na.rm=TRUE)Calculating weighted summary statistics
This is helpful is you want to compare your statistics to national data or to an average within your county or state . Why is it necessary to do it this way? If you average an outcome across all Census tracts, it will assign each tract an equal weight. That's an odd thing to do because some tracts have more people than others. When we use a weighted average, it essentially accounts for that.
Extremely brief example:
I have two tracts, one with a population of 5,000 where the poverty rate is 10%, and one where the population is 10,000 with a poverty rate of 20%.
If I take an unweighted average: , implying a 15% average rate. But that's obviously not right!
If I take a weighted average: , or a 16.7% rate. That seems better!
The SDMTools package previously described here likely won't work w/ your version of R! Try this instead: https://www.r-bloggers.com/2020/01/using-summarise_at-weighted-mean-tidyverse-approach/
Subset your variables
xxxxxxxxxx## restrict observations to Wisconsinwisconsin <- filter(atlas,state == 55)## restrict observations to Milwaukee Countymilwaukee <- filter(atlas,state == 55 & county == 079
How has inequality changed for countries in a particular region over time?
The World Bank data repository has excellent country-level data on inequality. You can download it directly and then import it into R, or you can use the WDI() package. If you go for the package, you will want to spend some time with the documentation to understand how to use it.
Explore trends in inequality in labor market outcomes (unemployment rates, labor force participation rates, wages) by demographic subgroups of your choosing.
Recall that you can import just about any Excel sheet or .csv file into R! That means all the resources you used in Unit 01 (FRED, BLS, IPUMS, etc) are at your disposal.
To really step things up, check out Ray's demo of a neat package that lets you easily access BLS data!
There is also some pretty easy-to-work with data embedded in a recent #TidyTuesday. See the description and links in the 08-01 PC for advanced users
This is fairily similar to the Capstone 01 prompt! If you do this, make sure you are doing a substantially different topic from Capstone 01.
If you have another idea, run it by me first! I'm pretty open.
To be suitable it must (a) have a clear connection to one of our topics in Units 1 or 2; (b) lend itself to analysis with data; (c) be substantially different from your Capstone 01 topic.
All the details are here! 👇 👇 👇
Specifications powerpoint pptx pdf
Slide deck as a PDF
Slide deck as .pptx
Knit output file (pdf/html/docx) that contains:
Data
Using a package: Skip this step, as you will call the data in your code!
Importing data: Include the raw .xlxs or .csv file
One way to look at your data is to plot a distribution of an outcome. You can do this with a histogram with one variable or multiple variables! You might also consider better ways of showing this, using a smoothed density plot or a violin plot.
Download this set of density examples: