Law School Predictor: The Thread" Forum

[nymets123]

The GPA calculator on the law school predictor is off by .01 (.01 higher than it really is).

is it really? My GPA on law school predictor and my LSAC GPA were exactly the same

[YCrevolution]

..

[nativedelta]

Ah--so two schools I got into should have been a straight out deny. And one school was only a weak consider.

Nice to beat the calculator, because I certainly agreed with its predictions.

[YCrevolution]

..

[thesealocust]

edit: n/m

[YCrevolution]

..

[thesealocust]

edit: n/m

[YCrevolution]

..

[Bronte]

If I knew how and knew it wouldn't be against their TOS, sucking LSN data would definitely be a good option.

This is something I'm very curious about as well. Where does LSC and hourumd get the raw LSN data that they use? I was expecting to find .xls versions on LSN, but that's definitely not happening. Do you think they sell it?

[YCrevolution]

..

[YCrevolution]

..

[CyLaw]

If I knew how and knew it wouldn't be against their TOS, sucking LSN data would definitely be a good option.

This is something I'm very curious about as well. Where does LSC and hourumd get the raw LSN data that they use? I was expecting to find .xls versions on LSN, but that's definitely not happening. Do you think they sell it?

At some point in the past (at least a couple years ago, I think), they used to release CSV data on LSN users. They've stopped doing that.

As for Hourumd and LSC, my guesses are:
A. They're using the old CSV files.
B. Someone good at web programming figured out a way to harvest them from LSN. This harvesting might be dynamic (real-time or close to it) or requires manual activation.
C. They asked/knew someone at LSN who gave them the data (I doubt it would be very recent just because my very limited experience with LSN tech support leads to me think they wouldn't be that helpful, at least not for free).

Through some quick examination, I discovered that you can get all the lsn data for a given school for a given cycle.

http://www.lawschoolnumbers.com/stats/a ... Cycle=0809

Will give you all the datapoints for georgetown (apparently indexed as 7 on lsn) for 2008-2009 in a nice xml format.

It would be trivial for someone to write scrapper program to get all the data points for all the schools.

I will not have time until this weekend to do this, but if you can wait tell then, I can do it. Since it is publicly accessible data of user created content and you are only using it to verify your formulas, I imagine it should be okay. Maybe one of the actual lawyers wants to comment on if it is fair use or not.

[thesealocust]

edit: n/m

[YCrevolution]

..

[YCrevolution]

..

[thesealocust]

edit: n/m

[Bronte]

Through some quick examination, I discovered that you can get all the lsn data for a given school for a given cycle.

http://www.lawschoolnumbers.com/stats/a ... Cycle=0809

Damn, how did you find this? Did you have to look at the source code? I couldn't find it through the regular interface.

Also, given how the data play out (and forgive me if I'm late to the party in terms of throwing out ideas) - wouldn't it make sense to just spit out what % chance LSP thinks you have rather than a couple of vague words? Or at least have that option - since I'm sure that's what is happening in the background anyway?

We were discussing this a few posts up, starting with PrayingforHYS's post.

[CyLaw]

Yeah I had to look at the webpage source code after I determined that the graphs were actually done in flash. In the source they had that URL as the argument to the flash object so I just typed it into my browser and out comes the XML. Definitly not a listed feature on LSN.

[Bronte]

Yeah I had to look at the webpage source code after I determined that the graphs were actually done in flash. In the source they had that URL as the argument to the flash object so I just typed it into my browser and out comes the XML. Definitly not a listed feature on LSN.

Right, I see it now. Hmm. I wonder how easily this could be imported into Excel. Do you have experience writing scripts that parse XML? I have a morbid fascination with recreating these prediction models...

Edit: It does in fact import into Excel no problem, not that anybody besides me cares.

[CyLaw]

Yeah, I was going to write a scrapper to get all the data and convert it to a usable format for YCrevolution this weekend. Speaking of, YCrevolution what format do you prefer, I can do whatever, but I guessed that the one you would want was a csv file with username, lsat, gpa, law school, result. Correct? I can also give it to you as a database or whatever format you need.

[YCrevolution]

..

[Bronte]

CSV will work well (basically anything I could import into Excel). Ideally, it'd be something along the lines of (order doesn't really matter)

University of Virginia,Accepted,170,3.85,UserName

Thanks so much!

Don't forget URM! The XML files don't contain a binary true/false for URM status, so it has to be extracted out of the username, where the string "URM" is listed in parentheses if true.

[YCrevolution]

..

[Bronte]

Good catch. ED and scholly money would be nice, too, but I'm not seeing anything that looks like those in the link.

Revised example: University of Virginia,Accepted,170,3.85,URMTrue/False,UserName

This can actually be easily added in Excel after the fact if CyLaw doesn't get to it. I'm using:

Code: Select all

=IF(ISNUMBER(FIND("URM",D2)), 1, 0)

Where "D2" is the username column. It returns "0" for non-URM, "1" for URM.

And, yeah, scholly/ED is not included as far as I can tell. LSN uses these XML files simply to generate the graphs, that's why the data isn't as rich as it could be.

[YCrevolution]

..