I’m upset that I’m not perfect and I’m also very tired.
The Invisible Developer asked me tonight if I had a list of all of the grants that I’d had funded in my career. For some reason, he thought I should have kept track of that. I told him that no one cared, not even me. The first few years, I would list in my c. v., “Over two million in funded projects.” “Over three million in funded projects.” Eventually, I felt like one of those McDonalds signs that keep changing, “Over 427 billion served.”
So, if I mentioned grant writing at all, I would just list a half-dozen or so funded projects. Really, once you’ve brought in over $7.5 million, people don’t care so much about the details. Dr. Erich Longie, who was president of Cankdeska Cikana Community College when I worked for them used to say we had gotten over $30 million. I honestly forget. I sat down and wrote whatever I could remember and came up with about $19 million, but I’m sure there are some from 15 or 20 years ago that I forgot. Of that $19 million, $15 million were grants I wrote with no help whatsoever. That is, I sat in front of computer, swore, wrote, added numbers, wrote some more, swore some more and in the end produced 100 or 200 pages that were good enough that someone gave the funding to run a program and pay people’s salaries for five more years. The other $4 million, I wrote large parts of but other people helped with budget or some other part.
I know I have forgotten a bunch, and I don’t even care to look. When I was trying to come up with a list, I saw one grant for $1.5 million in my list of examples and thought, “Oh yeah, I had totally forgotten about that.”
And yet, today I made a mistake on a grant and I kept thinking what an idiot I am.
I used to make a lot of money writing grants for people. There was one year when every single proposal I wrote got funded. I was the flavor of the month. Everyone wanted me to write grants for them. Of course, the first time I wrote one that didn’t get funded, the client was pretty upset. How could that happen?
In an average year, 85% of the proposals I wrote got funded. I’m not as great as that makes me sound. I was selective in what competitions I chose. If there wasn’t at least a one in seven chance of getting funded, I didn’t apply. When I started, my cut-off was one in five, but things have gotten more competitive. (That is, I would look at the number of proposals they funded last year and the number of applications they received and calculate the odds. Some competitions fund as few as 3% of the applicants. I wouldn’t bother with these.) Still, hitting 85% when only 14-20% get funded is a pretty good track record.
I don’t think the mistake that I made will keep the grant from getting funded. It was possible to submit a revision, since it was before the deadline, and I did that, on time. Still, I felt like an idiot.
Confession: I don’t really like grant writing
I’ve never liked grant writing and I’ve only ever met one person who did. He was very good at it but he’s retired now and probably nearly 80. It’s tedious work. You read 100+ pages of instructions, write 100 pages to fit a formula – Needs Assessment, Objectives, Project Design, Evaluation, Adequacy of Resources, Personnel. Fill in every box and bubble. Cite research to back up everything.
This is probably why I’ve never been very sympathetic when my children complained about their schoolwork being boring or hard. The fun, easy stuff we do for free. The boring or hard stuff, you need to pay. Grant writing is both boring AND hard. I did it for years because I was a widow with three young children and I needed the money. I’m grateful that I was able to support my children through private schools, good universities and the Olympics.
The only grants I write any more are for people I have worked with for years. Don’t call and ask me if I will write one for you because the answer is “No.”
No matter how many millions I hit, I feel terrible when I miss
Often, I’m writing grants for institutions where people are on soft money. That means, if the grant I write isn’t good enough, people lose their jobs. So, people somewhere else who did get the grant keep their jobs, but I don’t know THOSE people.
My point, and I do have one
I was going to write about PROC DATASETS today, but I wrote about this instead because it has been on my mind.
I think I have this in common with a lot of successful people – no matter how much money I bring in, how many good papers I write, no matter how many keynotes “knock ’em dead”, no matter how many grants get funded – if I slip once, I think I’m a failure.
Realistically, I know this is not true. If this particular grant doesn’t get funded, I’ll still have written tens of millions of dollars in successful proposals. This struck me the LAST time I had something not funded, over a year ago. I was feeling bad about it, and happened to be looking for something in a filing cabinet. (Yes, we have filing cabinets.) Going through those, I came upon file after file of data, reports, budget reports, from one funded project after another. It occurred to me that I had a LOT of successful projects over the years. It’s like this proposal I just finished. I think it was excellent work but there were one or two mistakes, and they weren’t even fatal mistakes (I hope!)
Here is my advice – successful people tend to immediately forget their successes and focus on the next challenge. That may be part of what makes them successful but it can also be a bit depressing if you forget the successes of the past when you are confronted with a failure in the present it looms unrealistically large. So, yes, fix your mistakes, learn from them, but also take some time every day to pat yourself on the back for the many, many mountains you’ve climbed in life.
Last week was very productively spent at Unite 2014 learning about all things Unity.
In case you are not into game development, Unity claims to be used by over a million game developers around the world. While I rather suspect those statistics are up their with Second Life and Twitter counting everyone who ever signed up for an account, there is no denying that one whole hell of a lot of people use Unity for game development, including us. I have to say all of my major objectives in attending were met.
The first thing I wanted to achieve was make a definite decision whether to go with Unity for the 2D game for the iPad that we are going to dive into next month. We have some artwork, a rough design, but we’re coming up on the first point of no return decision. Well, there’s always a return, but if we start with Unity and then switch to solution X it may take us quite a bit of time to re-tool.
The decision was, yes, we definitely want to use Unity. My concerns about performance on lower powered devices were addressed. First, I spoke to some helpful folks from Unity who pointed out that you can set your game’s graphic quality ranging from Fastest through Beautiful to Fantastic. Yeah, those are actually the last two settings. I also attended a session on tips for working with mobile devices that gave me some good ideas, like if we have character that has a sword, instead of having two images, a sword and a character, have that be drawn as one image.
Two other clinchers for unity were
the number of vendors with integrated add-on packages, everything from SpeedTree, which makes drawing trees fascinating to mixamo which offers a much simpler way for making 3-D animated characters. I was so impressed with mixamo that I texted one of our fabulous artists from the presentation, This is something we need to start using, and by we, I mean you, because we both know I suck at art.
The second thing I wanted to achieve was to get more familiar with Unity. That was achieved. I was able to follow the examples in the Training Day and do the Nightmares game, which was pretty fun. The next couple of days, in my spare time, I made another much simpler game from scratch for my grandchildren to play. It won’t win any awards for originality or anything else, but my Unity knowledge definitely spiked up in a week.
One reason I insist on going to events like this, even though people tell me that I am the CEO and should be doing CEO things, is that I would never, ever find 40 hours in a week just to learn if I stayed back in the office. I’ve written before about the Red Queen’s Race in technology, where you need to run as fast as you can to stay in the same place. I turned 56 last week and more opportunities are coming my way than ever before, which I attribute to refusing to equate age with stagnation.
No brogrammer culture in sight
Speaking of age – I usually go to conferences on statistics – the Joint Statistical Meetings, SAS Global Forum, etc. Sometimes I go to start-up events. This was my first game developer conference and I had heard horrid things about the game industry, that women are sexually harassed, assaulted, disrespected.
As far as horrid brogrammer culture – didn’t see it, and I looked. The demographics were overwhelmingly male, somewhere between 90-95%, I would guess. None of the sessions I attended had a female presenter. On the other hand, I didn’t submit a paper. I suggested it to The Invisible Developer and he didn’t want to do it, and I was too busy with everything else. We decided next year, for sure we would co-author one and submit it. Should be fun.
My point is, I don’t think they received hardly any submissions from women, just based on the number of women attending.
Despite all of the people who claim to have started coding in the womb and how much VCs supposedly drool over twenty-somethings, I saw about as many people under 20 as I did over 60. That’s based on me eye-balling it, I didn’t actually go around carding people. Given the amount of grey hair and balding, I’m going for the crowd was overwhelmingly in their thirties and forties.
While there were far fewer women than at statistics conferences, there were more than the zero African-Americans and Latinos you usually see at statistics events, although it was clear from the eavesdropping during the coffee breaks (I call it qualitative data collection) that many of these folks were actually from Latin America attending the conference. It was FAR more international than SAS Global Forum or JSM, even though both of those have a smattering of international folks.
As far as the whole sexual harassment, mansplaining, unwelcome thing – didn’t see it. Nada. Zip. Zilch. Every single person we met was nice, polite and interested in talking about game development. No one treated me like I was a second-class citizen and the only person who insisted on explaining stuff to me that I already knew was The Invisible Developer, but he has lots of other non-annoying traits that make up for it, so it doesn’t bother me.
It may be that I am old, plus I was there with my husband, so no one would bother me. However, I really did look, whether it was at cocktails in the evening, at lunch, during the coffee breaks, at the young women sitting around me in conference sessions – and I did not see a single hint of the kind of bad behavior I’ve been hearing about. I’m a small person and at this conference, I was just there to hang out and learn stuff, so I was wearing jeans and a hoodie most days, my point being, there wasn’t any reason people would be on their best behavior around me.
I’m not saying it doesn’t happen. I’m saying I didn’t see it happen here.
All I can say is — you should go to the next Unite conference. Learn stuff about game development and people will be nice to you. What more can you want? Well, if you want more, I should add that Seattle had some awesome restaurants.
If you want to go next year, do jump on it right away when you see it advertised though, because everything sold out – the conference, training day, nearby hotels.
Lately, I’ve been working on a report that uses eight datasets that all have the same problems with the usernames.
In addition to needing to remove every username that contained the word “test” or “intern” we also needed to delete specific names of the classroom teachers who had played the game. We needed to correct names that were misspelled.
Here are a few examples of a very long list of statements:
if username in("MSDMARIA","1ANSWERKEY","MSDELAPAZ","MSCARRINGTON") then delete ;
if username = “GRETBUFFALO” then username = “GREYBUFFALO” ;
else if username = “HALFHORES” then username = “HALFHORSE” ;
else if username =”TTCARLSON18TTCARLSON18″ then username = “TTCARLSON18” ;
These problems occurred in every dataset.
A second problem found when looking at the contents of each of the 8 datasets was that the username variable was not the same length in all of them, which could cause problems later when they were to be merged together or concatenated. Also, now that all of the usernames have been cleaned up, none should be over 12 characters in length.
Wouldn’t it be nice if there was a way to just get the first n characters of a string?
Enter our character function, substr, which returns a substring of a variable beginning at any position and for as many characters as you like. Problem solved.
newid = substr(username, 1, 12) ;
It seems pretty inefficient to write this set of statements eight times in eight different data sets. Also, next year we will have another eight data sets, and some will have these same students’ usernames and same problems. Wouldn’t it be a lot easier to have these statements in one place and add to the “fixnames.sas” file whenever we find a new problem?
So, now we have the write once, use anywhere solution of %INCLUDE.
What %INCLUDE does
The %INCLUDE statement references lines from an external file and processes them immediately. It has almost the exact same effect as if you had copied and pasted those lines write into your program. The exception that makes it “almost” is that a %INCLUDE statement must begin at a statement boundary. That is, it has to be either the first statement in your program or occur after a semi-colon ending a statement as in this example.
data studentsf ;
infile inf delimiter = “,” missover ;
attrib teacher length = $12. username length = $ 16. ;
input username $ age sex $ grade school $ teacher $ ;
%include “/courses/u_mine.edu1/wuss14/fixnames.sas” ;
Also, you need to think about it as if you had copied and pasted those lines into your program. Is it still valid code? Whenever using %INCLUDE, you should make sure the code runs in your program as expected, with no errors, before cutting it out and making it an external file.
To source or not to source
The default is not to show the statements that were in the included file. Generally, this is desirable. This is code you have already debugged and if you are using it multiple times (otherwise, why bother with the %INCLUDE), having the same 20 lines repeated 8 times in your log just makes it harder to debug.
Professors might want to use real data but hide all of the messy data handling from the students initially in fear they would run screaming for the door. I meant, professors might want to gradually introduce SAS statements and functions for data handling.
In either case, students could use the %INCLUDE statement as shown in the example above. To see the code include in your log is quite simple, just add a source2 option as shown.
%include “/courses/u_mine.edu1/wuss14/fixnames.sas” /source2 ;
It will be in your log as follows
NOTE: %INCLUDE (level 1) file “/courses/u_mine.edu1/wuss14/fixnames.sas is file “/courses/u_mine.edu1/wuss14/fixnames.sas.
419 +username = compress(upcase(username),”. “) ;
420 +if (index(username,”TEST”) > 0 or index(username,”INTERN”) > 0
and so on.
The + signs next to each statement denote it was in the included file.
If you want to know why I think it is so important for new SAS users to learn about the %INCLUDE statement, you should come to the Western Users of SAS Software conference in San Jose next month. Especially if you’re a student, you should come, because they cut you a really good deal.
If you’re not a student and you have a real professional job – well, then, you should be able to afford it. There will be funny hats, beer, coding and cookies. What more could one ask?
A few years ago, I was at the Western Uses of SAS Software conference and renowned statistician Stanley Azen played the piano and sang at the closing ceremony.
Briefly, very briefly, I considered beginning my presentation on 10 SAS steps to an annual report, by writing a song, These are a few of my favorite PROCs, and then singing it to the tune of “These are a few of my favorite things.”
This plan was dismissed a nanosecond later when I was reminded by The Perfect Jennifer that my singing bears an uncanny resemblance to the sound Beijing the cat used to make in the middle of the night when fighting with the cat next door.
To make up for my disappointment over my lack of musical rendition, I decided to do a few posts on my favorite PROCS, in no particular order. Today’s contestant is … drum roll please ….
Whenever possible, I try reading in the data using the IMPORT procedure, because, it is very simple and my goal in programming is not impress people with my brilliance – it is to get the job done with maximum efficiency and minimum effort.
As can be seen in the example below, there is no need to declare variable lengths, type or names. Only three statements are required.
PROC IMPORT OUT= work.studentsf DATAFILE= "/courses/u_mine.edu1/wuss14/Fish_students.csv" DBMS = CSV REPLACE;
GETNAMES = YES ;
This PROC IMPORT statement gives the location of the data file, specifies that its format is csv (comma separate values), the output file name is studentsf, in the work directory and that if the file specified in the OUT= option already exists, I want it to be replaced.
The second statement will cause SAS to get the variable names from the first row in the file. Since the variable names are in the first row of the file, the data begins in row 2.
Limitations of PROC IMPORT
As handy as it can be, PROC IMPORT has its limitations. Three we ran into in this project are:
- Excel files cannot be uploaded via FTP to the SAS server, so , no PROC IMPORT with Excel if you are using the SAS Web Editor,
- If the data that you want to import is a type that SAS does not support, PROC IMPORT attempts to convert the data, but that does not always work.
- For delimited files, the first 20 rows are used to determine the variable attributes. You can give a higher value for the number of rows scanned using the GUESSINGROWS statement, but you may have no idea what that higher value should be. For example, the first 300 rows may all have numbers and then the class that was records 301-324 has entered their grade as “4th” instead of the number 4.
Although PROC IMPORT is the first thing I always try, one of my pet peeves about instructors and textbooks is when it is the only thing they teach. It’s smart to try the simplest solution first. It’s dumb not to have a back up plan for the instances when that doesn’t work.
For more on what to do in those cases, you can come to WUSS in San Jose. Just a reminder – regular registration closes August 25. After that date, you’ll have to register on site.
Finishing up my second paper for WUSS next month and I have been thinking about the usefulness of character functions in a world where it sometimes seems like everyone is just put on this earth to irritate the hell out of me.
Take this problem, for example,
In analyzing the data for our games, we have all sorts of beta testers – teachers, staff, interns – who played the game but their data should be deleted from the datasets for the annual report. We asked them to use the word TEST in their username so it would be easy to pull them from the data. Some of them did and some apparently feel that I just say these things of exercise for my mouth.
There is also a problem with data entry errors. The subjects in this study were children in grades three through six and they frequently mistyped their usernames.
SAS has a wealth of character functions and this is a first opportunity to get to know and love four of them.
The UPCASE function, not surprisingly, changes the value of a variable to upper case. The COMPRESS function, if you give it only the variable as an argument, will remove blanks from a value. You can, however, include additional characters to remove. Since many of the students entered their names on some days as JohnDoe and others as John.Doe , we are removing both blanks and periods using the COMPRESS function, after we have converted them to upper case.
username = COMPRESS(UPCASE(username),'. ') ;
Then there is the INDEX function. Here is a general tip. Any time you find yourself thinking,
“Gee it would be nice if SAS did thing X”,
it is a pretty good bet that someone else thought the same idea and there is a function for it. The INDEX function is a perfect example of that. Our testers played the games many, many times and used usernames like “tester1”, “this.test”, “skippy the tester” or “intern7”.
“Wouldn’t it be nice if there was way to find out whether a given string appeared anywhere in a value?”
Enter the INDEX function, which does exactly that. This function is case-sensitive, but since we already converted the username to upper case above, that is no problem for us.
IF INDEX(username, “TEST”) > 0 or INDEX(username,”INTERN”) > 0 THEN DELETE ;
will do exactly what we want. The INDEX function returns a number that is the starting position in the string of the substring we are trying to find. So, in “skippy the tester”, the value is 12, in “tester1” it is 1. If the string is not found, the value is 0.
A problem I found when looking at the contents of each of the 8 datasets used for my research project was that the username variable was not the same length in all of them, which could cause problems later when they were to be merged together or concatenated. All of the usernames should have been a maximum of 12 characters but there were data entry problems when students would type mister_rogers instead of mr_rogers.
When the data are read in using PROC IMPORT, “For delimited files, the first 20 rows are scanned to determine the variable attributes. You can increase the number of rows scanned by using the GUESSINGROWS data source statement.”
Wouldn’t it be nice if there was a way to just get the first n characters of a string?
newid = SUBSTR(username, 1, 12) ;
will create a new variable with the first 12 characters of the username, now that we have gone and fixed the problems with it.
SAS is chock full of functions and options to make your life easier. If you are just beginning to work with SAS and you spend time working with messy data, you probably couldn’t spend your time much better than taking a few hours to read up on SAS character functions. In fact, I think for someone new to SAS, becoming familiar with a large number of all types of functions – character, statistical, date and time – is probably the fastest way to improve one’s productivity. (Ron Cody’s book, SAS Functions by Example, is a great resource). I’ve lost count of the number of times when reviewing a student’s program I’ve seen many lines of completely unnecessary code that could have been replaced by a SAS function – if the student only knew that it existed.
The second time I taught statistics, I supplemented the textbook with assignments using real data, and I have been doing it in the twenty-eight years since. The benefits seem so obvious to me that it’s hard to believe that everyone doesn’t do the same. The only explanation I can imagine is that they are not very good instructors or not very confident. You see, the problem with real data is you cannot predict exactly what the problems will be or what you will learn.
For example, the data I was planning on using for an upcoming class came from 8 tables from two different MySQL databases. Four datasets had been read into SAS in the prior year’s analysis and now four new files, exported as csv files were going to be read in.
Easy enough, right? This requires some SET statements and a PROC IMPORT, a MERGE statement and we’re good to go. What could go wrong?
Any time you find yourself asking that question you should do the mad scientist laugh like this – moo wha ha ha .
Here are some things that went wrong –
The PROC IMPORT did not work for some of the datasets. No problem, I replaced that with an INFILE statement and INPUT statement. It’s all good. They learned about FILENAME and file references and how to code an INPUT statement. Of course, being actual data, not all of the variables had the same length or type in every data set, so they learned about an ATTRIB statement to set attributes.
Reading in one data set just would not work, it has some special characters in it, like an obelus (which is the name for the divide symbol – ÷ now you know). Thanks to Bob Hull and Robert Howard’s PharmaSUG paper, I found the answer.
DATA sl_pre ;
SET mydata.pretest (ENCODING='ASCIIANY');
Every data set had some of the same problems – usernames with data entry errors that were then counted as another user, data from testers mixed in with the subjects. The logical solution was a %INCLUDE of the code to fix this.
In some data sets the grade variable was numeric and in others it was ‘numeric-ish’. I’m copywriting that term, by the way. We’ve all seen numeric-ish data. Grade is supposed to be a number and in 95% of the cases it is but in those other 5% they entered something like 3rd or 5th. The solution is here:
nugrade=compress(upcase(grade),'ABCDEFGHIJKLMNOPQRSTUVWXYZ ') + 0 ;
and then here
Data allstudentsents ;
set test1 ( rename =(nugrade= grade)) test2 ;
This gives me an opportunity to discuss two functions – COMPRESS and UPCASE, along with data set options in the SET statement.
I do start every class with back-of-the-book data because it is an easy introduction and since many students are anxious about statistics, it’s good to start with something simple where everyone can succeed. By the second week, though, we are into real life.
Not everyone teaches with real data because, I think, there are too many adjunct faculty members who get assigned a course the week before it starts and don’t have time to prepare. (I simply won’t teach a course on short notice.) There are too many faculty members who are teaching courses they don’t know well and reading the chapter a week ahead of the students.
Teaching with real, messy data isn’t easy, quick or predictable – which makes it perfect for showing students how statistics and programming really work.
I’m giving a paper on this at WUSS 14 in San Jose in September. If you haven’t registered for the conference, it’s not too late. I’ll post the code examples here this week so if you don’t go you can be depressed about what you are missing,
Hint: It’s math!
I was thinking about that again today. Unity is the game engine that claims to be used by over a million developers. That may be true – it seems like everyone I ran into at the Serious Play conference was using Unity, and we do, too.
Unity is great and there are an enormous number of assets you can purchase that make it easier to create games. That being said, at the first step in learning Unity, you are told that to locate your object in this three-dimensional space you set the X, Y and Z values. The default for these is the origin (0,0,0).
If you’re reading this blog, you’re probably the kind of person to whom that is immediately obvious. You’ve looked at charts a thousand times, you know exactly what an X and Y axis are, that the origin is the point at X=0, Y= 0 and it is not much effort at all for you to conceive of a Z axis and generalize from two-dimensional space to three dimensions.
When one of the most basic tutorials begins with discussing a plane, even if you didn’t remember much about geometry, you probably would recognize that as a two-dimensional space.
Concepts like scale and rotation depend on mathematics.
I’ve been trying to think of examples of programming that didn’t use much math. I did come up with one – you could create an application using PHP, MySQL, HTML and CSS to enter data into a database via forms and create simple reports. Not sure how many kids would be interested in that – I don’t find it all that interesting myself and I love programming.
My point is that if we don’t teach kids math, they are going to be limited in the types of coding they can do. Even areas like gaming, where you might think math isn’t so necessary, depend heavily on a level of mathematics that the average American student struggles with.
Now, if people who are promoting teaching kids to code see it as one way to motivate students to learn more math, then I think they might have some success.
On the other hand, overlooking the fact that students will need math is setting them up for failure.
So, that’s why the proposal I’m working on now is to develop games to teach students geometry, statistics, measurement and data. I hope that then, there will be a larger pool of young people prepared to learn to code.