*******************************
*** Firse Steps ***************
*******************************

*** Always set your working directory first
cd D:\Stata          // change your working directory (you should choose your own directory and enter the full path here)

*** Use a log file to keep track of whatever you have done
log using log1,replace
// You can only keep one log file open, 
// so opening another one will give you an error message:
log using log2,replace
// The command will work if you close the existing one.
// So you should close all existing logs:
log close _all
// now let's open the first log file again, and apecify the "text" option to save it in txt format
log using week1.txt,text replace 
// Let's also create a command log:
cmdlog using cmdweek1.txt,replace 
// We will learn what it is later.
*******************************
*** View Data
*******************************
*** Open a Dataset
sysuse auto,clear    // clear existing dataset in memory
// Check what the variables and properties panels look like
*** Think of a few ways how you would like to explore the data:
*** Most directly way is to simply take a look of the data:
browse
// Structure of the data
// Explain color, missing, categorial data
// Note that you can only read the data without editting it.
// In order to edit the data, type "edit" instead:
edit
// I strongly discourage you from editting data this way.
*** To know more details about the data (structure)
describe
// clear more
// elements displayed
*** Summary statistics
summarize
// why "make" has no observations
// rep78 have less obs, because of missing values
*** To learn the contents of the data in detail
codebook make // unite values
codebook foreign // value labels
codebook rep78 // missings
*** To learn more about the missing data
browse if missing(rep78)
list make if missing(rep78)
*** More summary statistics, e.g., median
summarize,detail // notice that mean is very small
browse if price>13000
*** Learn the composition of foreign/domestic and repairing records
tabulate foreign
tabulate rep78
tabulate foreign rep78,row 
*** To compare gas mileages between foreign and domestic cars:
*** Think about the potential ways to do this
*** Method 1
summarize mpg if foreign==0
summarize mpg if foreign==1
// equal sign
*** Method 2
by foreign,sort: summarize mpg
*** Method 3
tabulate foreign, summarize(mpg)

*** Hypothesis testing
ttest mpg,by(foreign)

*** Relationships between mpg and weight
correlate mpg weight
by foreign,sort: correlate mpg weight
correlate mpg weight length turn displacement

*** Graph data
twoway (scatter mpg weight)
twoway (scatter mpg weight),by(foreign, total)

*** Model fitting
generate wtsq=weight^2
regress mpg weight wtsq foreign
predict mpghat
twoway (scatter mpg weight) (line mpghat weight,sort),by(foreign)
generate gp100m=100/mpg
label variable gp100m "Gallons per 100 miles"
twoway (scatter gp100m weight),by(foreign,total)
regress gp100m weight foreign

*** Close and save the log files before you go
log close _all
cmdlog close