// analysis using stata workshop, econ311
// kbott | instructional technology services | reed college
// analysis questions? feedback? kbott@reed.edu
// general data questions? contact the data@reed team at data@reed.edu


//////////////////// INTRO AND BASICS
** give Stata permission to not pause at the end of every results screen.
. set more off
// preloaded datasets, useful for training/learning
. sysuse dir
** select census dataset
. sysuse auto

** summarize full dataset or one variable (# obs, mean, stddev, min, max)
. summarize
** ...one variable
. sum rep78

** describe full dataset 
. describe 
** or one variable (var name, storage type, disp format value label variable label **
. d rep78

** step one is always LOOK AT YOUR DATA!
** open data browser
. browse
* note: can also just type " br "
// . br

** codebook, full dataset or one variable (type, range, units, unique, missing, mean, stdev, %tiles)
. codebook
. codebook rep78

** clear out data in memory (good practice; not required for system data)
. clear
. sysuse auto

** tabulate = variable, frequency, percent, cumulative %
** by variable
. tabulate foreign
** in pairs of variables
. tabulate foreign rep78
** hint for homework3
. tabulate make rep78

// KBOTT'S FIRST-GLANCE TOOLBOX
// .sum
// .codebook
// .d
// .tab
// .inspect
// .browse
// .list (in, if)

** browse = view data in data browser
. browse if foreign == 1
. browse if foreign != 1
** another way of stating this
// browse if foreign ~= 1
** & joins multiple
. browse if mpg > 5 & mpg < 20
** view subsets by range of values using in
. browse make mpg in 1/10

// ALTERING DATA -- these do exactly what they sound like. check help doc for syntax
//. sort
//. drop
//. keep
//. replace
//. gen
//. egen
** gen and egen have some of the same functionality egen = extended generate 

/////////////////////////////////// DATA ANALYSIS
/////////// VISUALIZE CAR REPAIR DATA (stock dataset) -- HISTOGRAMS
. sysuse auto
. hist price, freq
. hist price, freq bin(5)
. hist price, freq bin(15)

. hist price if foreign==1, freq bin(15)

/////////// VISUALIZE CAR REPAIR DATA (stock dataset) -- SCATTERPLOTS
. twoway scatter mpg weight
. twoway scatter mpg weight || lfit mpg weight
// ah, but you can be lazier in those commands...
. scatter mpg weight
. scatter mpg weight || lfit mpg weight

////////// ANALYZE CAR REPAIR DATA (stock dataset)
** correlation
. correlate mpg weight
** regression 
. regress mpg weight
** regression -- can also do by specific groups OR for subsets
. by foreign: regress mpg weight
. regress mpg weight if foreign==1

//////// HOMEWORK #3

// measures of central tendency + dispersion
//// summarize
//// tabulate
/// ... other commands?

// visualization
//// scatter

// analysis
//// regress
//// predict