Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
Damian Betebenner
@dbetebenner
OH (Ohio)
Damian Betebenner
@dbetebenner
I’m working on a bug of my own that caused a fail in testSGP(1)
Damian Betebenner
@dbetebenner
Adam, How long will it take to modify the cutscores for RLI so that we duplicate all the .3 cutscores as .4 cutscores
adamvi
@adamvi
Shouldn’t take too long. I’ll knock that out right now since I know the code. Do you have any other changes to SGPstateData to commit before I do that?
Damian Betebenner
@dbetebenner
Yes, I’m pulling in now.
Damian Betebenner
@dbetebenner
Thanks!!
adamvi
@adamvi
No problem! Just pulled into SGP master as well
Damian Betebenner
@dbetebenner
Man, this Arizona bug is just killing me. I can reproduce it and have Demonstration_SGP at the exact same place and I get an error with AZ but not with DEMO.
adamvi
@adamvi
That’s really bizarre. It must be something DEEP inside the data...
Damian Betebenner
@dbetebenner
Or (I’m afraid) something obvious that I’m not seeing. Whenever I start blaming DEEP stuff, I’m usually missing something much easier
adamvi
@adamvi
I debugged studentGrowthPlot issue and testSGP(1) passed. I’m going to pull my changes there and in SGPstateData into the masters.
Looking back on that error when it came up with Georgia, it looks like the problem was caused by duplicate cases present in tmp_object_1. You might check for that?
Damian Betebenner
@dbetebenner
Yep, there are duplicates!!!!
Which makes complete sense given that multiple cases are returned.
Damian Betebenner
@dbetebenner
Go ahead and pull your changes in whenever you can
Damian Betebenner
@dbetebenner
Thanks
Damian Betebenner
@dbetebenner
FYI. data.table now on GitHub https://github.com/Rdatatable
Damian Betebenner
@dbetebenner
Got the same error again for testSGP(1). Did you install off of GitHub before testing?
adamvi
@adamvi
Yeah and I tested both the master branch and mine. Both passed...
Damian Betebenner
@dbetebenner
This is pretty bizarre then that we’re getting two different results. My branch is showing that is it up to date with the master.
adamvi
@adamvi
Let me try testing your branch now to see if I get the same error too. Very bizarre...
Damian Betebenner
@dbetebenner
This issue (at least on my broken piece) is that there are no values in CONTENT_AREA_LABELS. They are all NA
Damian Betebenner
@dbetebenner
It may have something to do with a student who has only one content area and not another. I may have just got “lucky” that my random sample included such a student
adamvi
@adamvi
Must be something like that. Your branch passed too. I guess I’ll add in a conditional statement there that only executes the strsplit if its non-NULL and makes any other adjustments/assignments needed if not.
Damian Betebenner
@dbetebenner
I’m pretty sure it happens because of a student with a score on one test and not on another. They have a missing (NA) value for CONTENT_AREA_LABEL.2013_2014 that causes content_area.labels to not get set appropriately.
I’m testing my fix now
I think it’s the o with the accent on it that is causing the R warning. There must be some escape sequence for that character.
adamvi
@adamvi
\uF3 works. Changed the line to "Nivel de Capacitaci\uF3n” and that passes R CMD check (as cran) now.
Damian Betebenner
@dbetebenner
Nice
Damian Betebenner
@dbetebenner
Can you pull in your changes to the main
adamvi
@adamvi
Yes will do that now
Damian Betebenner
@dbetebenner
On Southwest. Let’s see how their internet runs compared to Delta and Jetblue
Damian Betebenner
@dbetebenner
Pretty decent! Am able to connect to EC2 without much lag
Damian Betebenner
@dbetebenner
testSGP(1) passed!
adamvi
@adamvi
Sweet! I’ve never had much luck with SWA internet, but glad you got a good connection!
Damian Betebenner
@dbetebenner
Latency is the big deal with ssh connections and it has pretty low latency. I don’t think data throughput is that great
adamvi
@adamvi
I see. Yeah I think I’ve tried to install from github before and the download time was terrible. Next time I’ll have to just ssh into my workstation!
Damian Betebenner
@dbetebenner
Getting a BIG 10 kbs download on a 900 megabyte file :smile:
adamvi
@adamvi
Dang! Fancy! Don’t go using the whole pipeline now - save some for the others ;)
Damian Betebenner
@dbetebenner
We’ve joined the ranks of the cool kids. Just put travis integration into my SGP repository!
Damian Betebenner
@dbetebenner
Adam, just pulled in my changes to the master branch after passing testSGP(1)
adamvi
@adamvi
Perfect - thank you! I’ll pull my changes in too once I’ve finished testing the other TCRP analyses.
Damian Betebenner
@dbetebenner
Pulled your SGPstateData changes down with no status changes to my repository (I think that was as planned)
Damian Betebenner
@dbetebenner
Adam, just added testSGP(6) to testSGP (5 is still a work in progress) to do a really basic check on baseline coefficient matrix construction and baselineSGP calculation. I don’t believe we had anything to just test that out.
adamvi
@adamvi
I was thinking we needed something like that when I was constructing new matrices the other day. Thanks for putting that together!
adamvi
@adamvi
I pulled my summarizeSGP changes in from utah. I still need to work on the SQLite solution for that, but the bug is at least fixed. This also pulled in the added ability to produce PNG bubble plots (although I still need to update the documentation for that too).
Damian Betebenner
@dbetebenner
Thanks.
Damian Betebenner
@dbetebenner
Also, since we are approaching a transition for most states, I’m contemplating just “unplugging” the baseline matrices for states since we won’t be using them.
Damian Betebenner
@dbetebenner
I’ve unplugged the matrices in SGPstateData for many states