Getting Help

Loading...

From the course by Johns Hopkins University

R Programming

11471 ratings

Johns Hopkins University

R Programming

11471 ratings

Course 2 of 10 in the Specialization Data Science

In this course you will learn how to program in R and how to use R for effective data analysis. You will learn how to install and configure software necessary for a statistical programming environment and describe generic programming language concepts as they are implemented in a high-level statistical language. The course covers practical issues in statistical computing which includes programming in R, reading data into R, accessing R packages, writing R functions, debugging, profiling R code, and organizing and commenting R code. Topics in statistical data analysis will provide working examples.

From the lesson

Week 1: Background, Getting Started, and Nuts & Bolts

This week covers the basics to get you started up with R. The Background Materials lesson contains information about course mechanics and some videos on installing R. The Week 1 videos cover the history of R and S, go over the basic data types in R, and describe the functions for reading and writing data. I recommend that you watch the videos in the listed order, but watching the videos out of order isn't going to ruin the story.

Meet the Instructors

Roger D. Peng, PhD
Associate Professor, Biostatistics
Bloomberg School of Public Health
Jeff Leek, PhD
Associate Professor, Biostatistics
Bloomberg School of Public Health
Brian Caffo, PhD
Professor, Biostatistics
Bloomberg School of Public Health

This video is about how to get help when learning to use R and taking this course.

Because of the size of the course and the number of students that are enrolled,

it's going to be difficult to be asking lots of questions on a one on one basis.

And so we're going to have to resort to a few other tools to get questions answered,

in particular, the discussion boards and generally through email.

And so with a certain type of way to ask questions that I hope to maximize

the chance of you getting the right answer or the answer that you're looking for.

And so, the main thing to remember is that asking questions via email is

a little bit different from asking questions in person.

You don't necessarily know that the people on the other side, the people that you're

asking, for example, on a discussion board or on a mailing list, you

don't necessarily know that they have the same background information that you have.

Furthermore, they may not know you personally and

so may not know kind of what you mean when you say certain types of things.

So that's kind of important to keep in mind when you're emailing questions

as opposed to when you're talking to someone in person.

Keep in mind, of course, that elder people are very busy, and their time is limited,

and although they may be willing to help you by answering a question,

they may only have a certain amount of time to devote to answering that question.

Now I am here, of course, as the instructor, to help you in all

circumstances, but furthermore, I may not be able to answer all possible questions.

And so you're going to want to use the resources that you have available to you

in this course.

So, in your search for answers, there are a variety of things that you can do

on your own before venturing off to ask other people for the answer.

So, if you're going to be emailing a question to a forum or to a mailing list,

it's important that you search the archives of that forum for the answers.

So it's possible, and depending on the size of the forum,

almost very likely that someone has asked the same question that you're asking.

And if someone else has asked that question, and it has been answered,

then the answer is going to be in the archive of that forum.

So if the answer's already there, you've saved everyone a lot of time, including

yourself, if you search the archives for the forum and just find the answer.

Of course, the web is very large and has many answers, and

your first reaction when you have a question is to search the web.

For given the type of program you're using, for example, here we're using R,

there are many manuals that are available, and many answers may exist in the manual.

There's a frequently asked question,

or a FAQ that's on the R website that you can look for that contains many

questions that are commonly come up on the mailing lists and on the forums.

Another thing that you might want to try to do before go

venturing out to ask people for the answer is to play around with the problem and

try to find the answer by inspecting or experimentations.

Maybe if you have a function that's not working right, maybe change the inputs and

see if the outputs change or if the error message changes.

If you're lucky enough to have a skilled friend who knows something about R,

you can ask them personally, and it's usually easier to ask a person

on a one on one basis than to email a group of people in a forum.

Furthermore, and lastly, if you're a programmer,

you maybe able to find the answers you're looking for by reading the source code.

So all the things in the previous slide are useful things to do

on your own before venturing out to ask people questions.

However, if you don't find the answer, it's important to let

other people know that you did all those other things on the previous slide.

Because if, for example, the answer in the documentation for the programs, for

example, it's in one of the R manuals that someone who knows

the answer will usually respond by saying read the documentation or read the manual.

And then you just wasted one round of email.

because, you'll have to respond saying, I did read the manual, and

I didn't find the answer there.

And so, letting people know that you've done your homework, and

you've looked in a variety of places is very useful, and it saves a lot of time.

So, here's a very simple example of what might happen as you're using R.

So, here I'm loading the data source package, and

then I'm going to load the air quality data source from that package.

And then I'm going to want to run the correlation function

on this air quality data frame.

So immediately, I get an error,

it says error in cor(airquality) : missing observations in cov/cor.

So you might be wondering well, why am I getting this error, what does it mean?

So the first thing you can do is go to Google.

And in many circumstances, Google's going to be your friend.

And this is no different when you're learning a new programming language or

when you're learning R.

And so, the easiest thing to do is to take the error message that you get and

literally cut and paste it into the search box for Google.

When I search on that, I get a number of results.

And particularly, this third result looks very promising.

It looks like someone asked this question on the R help mailing list, and

so maybe we're clicking on that to see if someone replied with the answer.

So, when you ask a questions on a mailing list, so

assuming that Google wasn't able to help you out,

if you're going to ask a question on the discussion board or on the mailing list,

there are a couple things that you need to think about before you ask that question.

First, is it possible to reproduce your problem?

So, if someone else can reproduce your problem, it makes it a lot easier for

that other person to figure out what the solution is going to be.

And so if you can provide some code or a very simple example that will reproduce

your problem, this will be enormously useful to everyone else involved.

And if you don't do this, typically, the first response that you'll get will be,

can you please provide a reproducible example?

Second, it's important to understand what you expect the output to be,

because if your expectation is wrong, then of course, it may or

may not be an error, depending on what your expectation should be.

So what you expect the output to be will indicate kind of what the nature of

the error and what needs to be solved.

And then given your expectation, you need to say, well, what do you see instead?

So what was the thing that was unexpected that gave you the question?

Other information that's important to specify when you're asking a question

is the version of the product you're using, so for example,

what version of R are you using?

What version of the R packages you're using,

if it's specific to a given package?

Because often there may be legitimate bugs in older versions of R or R packages,

and your problem might be solvable if you just upgrade to the latest version.

So if you're using the latest version of R, it's important to mention that.

Sometimes it's important to know what operating system you're using, so

whether you're using a Mac, or Windows, or Linux, or some other UNIX machine.

Some problems can be traced to the type of operating system that you're using.

And depending on the question that you have,

there may be additional information that you need to provide.

So when you send an email to a forum or to a discussion board,

it's important to get as much useful information in there as possible.

And this includes the subject line for the email.

So there's a couple of examples of subject lines that could range in usefulness.

So the first one is probably the least useful, it just says Help!

Can't fit linear model!

So here, there's very little information here.

All I know is that there's some problem with linear models.

I don't know anything else about what the user's problem is.

So the second version is much better, it tells me that on R version 2.15.0,

the lm() function produces a seg fault,

which meaning that R crashes, with a large data frame.

And furthermore, it says I'm using Mac OS 10.6.3.

So here, I've got the operating system,

the version of the operating system, I've got the version of R,

I've got what function I'm using, and I've got a summary of what actually happened.

So just a little bit smarter than that would just be to reformat the message, so

that I specify what version of R I'm using, the function,

and then the version of the operating system.

And then so that gives me the context, and then after that, I can say

what the problem was, which in this case, was a seg fault on a large data frame.

So here, the important details are right away put in

the subject header before I even get to the body of the message.

So, a couple things that you definitely want to do when you're asking a question

on a forum or a mailing list, first is to describe the goal, not the steps.

So you may have many, many steps that you're going through, and

maybe one of those steps is causing a problem.

It's useful for other people to know what the bigger picture is in terms of what

you're trying to do, because for example, they might have a better idea about how to

go about achieving that goal, which may be faster or simpler, and

may work around whatever problem you're having.

So, describe the ultimate goal, and then talk about what the problems are.

And don't just narrow it down to the one little step that you're having

a problem with.

Be explicit about your questions.

So remember, provide details about what you're trying to do.

And you have to provide the minimum amount of information necessary, so

not the maximum amount of information, the minimum amount of information.

And so it's common to see on some mailing lists posts that lots of

output is produced, and that's not very helpful.

because volume doesn't really help you in terms of diagnosing the problem.

You need to know exactly, to narrow down kind of where the problem is going to be.

So, and of course a couple of things, being courteous never hurts anyone.

And promoting civility on mailing lists is always a nice thing.

And if you find the solution later on, it's useful for

everyone else in the community if you follow up with the solution and

explain kind of what the problem was and how you solved it.

A couple of things you definitely do not want to do when posting to a forum,

you don't want to claim that you've found a bug.

This happens all the time, and usually, I'd say 99 times out of 100,

it's not a bug, and it's just a misunderstanding about what

should have happened, so a mistake in the expectations of the user.

Grovelling is a substitute for doing your homework,

that's not usually looked well upon, and you definitely shouldn't do that.

Definitely don't want to post homework questions on the mailing list or forums,

and the reason is because people who write the homework questions are reading those

mailing lists and will be able to identify all homework questions without a doubt.

So we've seen them all,

don't bother trying to get the answers to your homework on mailing lists.

Don't ever email multiple mailing lists at once.

So this is a little bit annoying, because people will subscribe to different

mailing lists and will be getting your message more than one time.

It's important to figure out which mailing list is the most appropriate mailing list

for your question and then send the message to that mailing list or forum.

And then lastly, don't ask others to debug your code without giving some sort of hint

as to what the problem might be.

So it's very difficult when a person posts a long listing of code and

says there's a problem in here somewhere, I don't know where, please help.

It's better to kind of specify where you think the problem is and

what you're trying to do, so that everyone can save some time.

So this is just a very brief case study on using a recent post to

the R-devel mailing list.

So, the R-devel mailing list is an email list for

people who are doing development work in R, so either they're developing packages,

or they're making modifications to the R source code itself.

And so

the subject was, so you can see from the subject that it's going to be a problem.

The subject is large dataset- confused.

So right away, you know that there's not that much information here to go on.

It's not clear what the problem is.

In the message, it says, I'm trying to load a dataset into R, but

I'm completely lost.

This is probably due mostly to the fact that I'm a complete R noob, but

it's got me stuck in a research project.

So you have to ask yourself,

what do I know about this person's problem from reading this message?

And the truth is very little, and so the response was somewhat predictable.

Basically, the first person response said yes, you are lost.

And then there is a pointer to the posting guide, which everyone should read before

sending an email to the mailing list, and then also a list of manuals.

And so, here, you can see that one round of email is immediately wasted,

because probably the answer to the question was in one of the manuals,

and so, the user didn't specify whether he or she had already done that.

So, in terms of what went wrong with this little exchange, first, of course,

the question was sent to simply the wrong mailing list.

The R-devel mailing list is for development questions and for

more sophisticated programmers.

It's not really a mailing list for questions for,

as this person stated, noobs.

So that question really should have gone to the R-help mailing list,

where it would have been better received.

However, in addition to that, the email subject was very vague,

it was not clear what the problem was.

The question itself was very vague.

There was no reproducible example there.

There was not possible for

other people who are reading the email to reproduce what that person's problem was.

And there was no evidence that any effort was made to solve the problem.

So there is no evidence that they searched the web or checked the manuals or

experimented with the problem, or anything like that, or even looked at the forums.

And so, the end result was a complete recipe for disaster, and

it's likely that this person did not get the answer to their question.

So a couple places to turn for this course, first of all,

the class discussion board I think will be the most useful because your fellow

students can help you out there, and I can respond to you on the discussion board.

Outside of class, there's the R help mailing list, which I just described.

And you can post to this mailing list as well, and it's useful to, of course,

follow all the rules that we talked about just now.

And depending on what other products you might be working on,

there are other project-specific mailing lists for other types of software.

And so this talk was inspired by Eric Raymond's posting called,

How to ask questions the smart way, and I encourage you to read that.

It's much longer, it has a lot of other useful tips.

Explore our Catalog

Join for free and get personalized recommendations, updates and offers.

Coursera provides universal access to the world’s best education, partnering with top universities and organizations to offer courses online.

© 2018 Coursera Inc. All rights reserved.

Download on the App Store

Get it on Google Play