Inspired by the new era of data journalism (Simon Rogers's Datablog on the Guardian is a perfect example) and Paul Bradshaw's collective investigation journalism http://helpmeinvestigate.com/, I decided to pour out my gibberish thoughts here. Don't take it seriously, just look at it as afternoon brain snacks.
Wednesday, August 17, 2011
Saturday, August 6, 2011
News Pick- International Higher Education
Below is my pick from related education/higheredu/international edu. Check it out.
Thursday, July 14, 2011
Hand-Flip Quantitative Analysis? Mm...Maybe
Quantitative analysis looks intimating with those formulas, algebra, calculus....But after all, we still call it Applied Econometrics. That means it is still usable, doable and interpretable.
Recall one of my favourite professors said in class that suppose you get good-quality data( I mean it, really structured), running a model is just hand-flip thing. The crux is how you interpret the results, namely the beta coefficents, and how you are going to use it to change the status quo.
Notes on Feb. 2nd, 2013: Dabbleboard seems to shut down for good. So the flow chart below is no longer available.
Recall one of my favourite professors said in class that suppose you get good-quality data( I mean it, really structured), running a model is just hand-flip thing. The crux is how you interpret the results, namely the beta coefficents, and how you are going to use it to change the status quo.
Notes on Feb. 2nd, 2013: Dabbleboard seems to shut down for good. So the flow chart below is no longer available.
Tuesday, July 12, 2011
Let's See Who Are Those International Students In The U.S.? Data From 2000-2009
Click on the bar and see the number of international students of that country. Enjoy~
Sunday, July 10, 2011
Friday, April 1, 2011
R: recode levels (missing value) to NA
R load dataset from spss and assign value to missing value. Also in my original dataset, they have coded missing value with a meaningful -8 -9. I hate it! After marathon search in online forum of R topic, I digged out my way to recode levels. Here are the solutions:
>attach(dataset)
> summary(variable)
>table(variable)
>table(as.numeric(variable))
>levels(variable) ##if that s a numeric factor, it will have error message. command above just to have idea of how many NA and what value of level to recode.
>levels(variable) [level value assigned] <- NA
#to check
>sum(is.na(variable))
# it tells you how many are missing, or
>table(as.numeric(variable))
#or
> levels(variable)
## NA level should appear then.
Note: if type:
> variable[1:4] <- NA ## it means assign NA to the 1st to 4th row under this column(variable)
> variable[3] <- NA ## Likewise, assign NA to the 3rd row under the variable
It is little bit confusing but finally made it!
>attach(dataset)
> summary(variable)
>table(variable)
>table(as.numeric(variable))
>levels(variable) ##if that s a numeric factor, it will have error message. command above just to have idea of how many NA and what value of level to recode.
>levels(variable) [level value assigned] <- NA
#to check
>sum(is.na(variable))
# it tells you how many are missing, or
>table(as.numeric(variable))
#or
> levels(variable)
## NA level should appear then.
missing value for numeric variables:
> sum(is.na(BYSES2))
[1] 0
> BYSES2[BYSES2==-8] <- NA
> sum(is.na(BYSES2))
[1] 305
> BYSES2[BYSES2==-4] <- NA
> sum(is.na(BYSES2))
[1] 953
> describe(BYSES2)
Note: if type:
> variable[1:4] <- NA ## it means assign NA to the 1st to 4th row under this column(variable)
> variable[3] <- NA ## Likewise, assign NA to the 3rd row under the variable
It is little bit confusing but finally made it!
Subscribe to:
Posts (Atom)