An R learning project (feel free to learn with me)

[trinker]

I'm thinking then that using ggplot2 for the google plots doesn't give me any advantage over what you've done here? Would you agree? (PS I tried your code for addresses in my area, great looking map and accurate.)

[bryangoodrich]

Don't know. Not entirely sure what the ggplot method was accomplishing. For simplicity, I'm trying to understand the functions that case study uses. For instance, it manipulates the graphic itself with that image package it uses. Frankly, this is probably because it is trying to transfer over what the RgoogleMaps plotting function does to the ggplot plotting structure. That requires some tweaking. The difference, I would suspect, is that the RgoogleMaps is limited in what you can do compared to what using other libraries can offer, but that takes work. For instance, if you wanted to generate the heat maps and topomaps, that the case study does, you're gonna have to be able to plot those lines and colors onto the map, which requires that your use of lines and points and other functions are in the same parameter space as your map canvas. This is something brought up in the RgoogleMaps vignette, if I recall correctly, but it references using another package, also.

For a simple pin map (plotting points-to-locations), there's no reason to use all the bells and whistles when the one package handles it all. But you will notice that RgoogleMaps uses the png library. I assume what it does is very simple: read an image raster in as a grid. It is then doing something similar to what the case study did using its own image processing. Frankly, I'd like to see this done in a more generic fashion that can be more easily translated to base-r. I'd also understand more what is going on!

[trinker]

I have been pretty busy the past couple of days but am moving forward on learning with this project. Earlier I had discussed wanting to read in an access file but thought I'd put it off. Well I discovered something interesting in this data set (a bi modal test distribution in counties that I believe have a higher poverty rate). I need to confirm the poverty thing and thus need some data to support this. This MS access data (LINK) is in a zip file so I am trying to read save it into the working directory using R and then read in the pieces I want using the RODBC package. I'm patterning this off of a post (LINK) bryangoodrich posed in the the R TIL thread. I'm running into some problems right now and not sure why. Any help is appreciated:

temp <- tempfile()
download.file("https://reportcards.nysed.gov/zip/SRC2010.zip", temp)
data <- read.table(unz(temp, "a1.dat"))
unlink(temp)

The 2nd step results in the following error:

> download.file("https://reportcards.nysed.gov/zip/SRC2010.zip", temp)
Error in download.file("https://reportcards.nysed.gov/zip/SRC2010.zip",  : 
  unsupported URL scheme

[Dason]

I don't believe download.file works with https.

[trinker]

Thanks Dason. I just noticed it was an https. Any other ideas or is manually downloading it the only way? RCurl maybe?

[Dason]

I'm pretty sure RCurl should be able to handle it.

[trinker]

Uggg! I'm getting an error that I got before but I can't remember what the error means or how to deal with it. I looked at RCurl's help manual and tried url.exists which R tells me it doesn't why? Is the is the reason for the error? Maybe no https will exist.

require(RCurl)
NYSdemo <- getURL("https://reportcards.nysed.gov/zip/SRC2010.zip")

#GIVES THIS ERROR
> NYSdemo <- try(getURL("https://reportcards.nysed.gov/zip/SRC2010.zip"))
Error in function (type, msg, asError = TRUE)  : 
  SSL certificate problem, verify that the CA cert is OK. Details:
error:14090086:SSL routines:SSL3_GET_SERVER_CERTIFICATE:certificate verify failed

> url.exists("https://reportcards.nysed.gov/zip/SRC2010.zip")
[1] FALSE

[Lazar]

I can get past the first error with :

download.file(url="http://curl.haxx.se/ca/cacert.pem", destfile="cacert.pem")

NYSdemo <- getURL("https://reportcards.nysed.gov/zip/SRC2010.zip",cainfo = "cacert.pem")

But then this creates a new error.

Error in curlPerform(curl = curl, .opts = opts, .encoding = .encoding)

[trinker]

I think I'm going to put this up on SO and see if any of the brains there can help. I'll link it back here.

[Dason]

On SO I can downvote you for being a raptor.

[trinker]

Oh great now they're going to know you're crazy there too. They had thought you somewhat normal before.

[Lazar]

I might change you post at SO a little trinker because that errer you got is a very common one that can be solved the way I showed or by including ssl.verifypeer = FALSE in the get URL call. The problem you have I think is different so you might not get the most useful feedback.

[Dason]

Oh great now they're going to know you're crazy there too. They had thought you somewhat normal before.

I don't need to leave a comment like:
(-1) For being raptor.

But you would know why. Although I'm not actually going to do that.

[trinker]

Lazar any ideas of a title? Suggested edits?

[Lazar]

Ok try the following

download.file("https://reportcards.nysed.gov/zip/SRC2010.zip", destfile="test.zip", mode="wb")

It started downloading something put the file is large so I stopped half way through.

EDIT: I have now tried this on two windows systems and it is working fine.