Med school data web app - calculate your chances and pick schools

paul411

ANES
7+ Year Member
May 27, 2010
1,616
33
Florida
Status
Resident [Any Field]
I've created a web app (from the data in the SDN-used created spreadsheet) to help people applying to med schools pick their schools. The app allows you to input your stats (GPA, MCAT) and it calculates your chances of getting into all the medical schools based on the academic stats. It also information about in-state vs. out-of-state numbers and costs. All the data is sourced from school websites and publicly available data.

You can access it here: Medical School Data App at http://paul411.github.io

It's in its infancy. I'll add more data and options (e.g. comparing schools) in the future. Let me know if you think of any suggestions and/or improvements.
 

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
I've created a web app (from the data in the SDN-used created spreadsheet) to help people applying to med schools pick their schools. The app allows you to input your stats (GPA, MCAT) and it calculates your chances of getting into all the medical schools based on the academic stats. It also information about in-state vs. out-of-state numbers and costs. All the data is sourced from school websites and publicly available data.

You can access it here: Medical School Data App at http://paul411.github.io

It's in its infancy. I'll add more data and options (e.g. comparing schools) in the future. Let me know if you think of any suggestions and/or improvements.
You may want to clarify what the Academic Index is and how it is calculated, it's quite hard to decipher off the SDN page

Excellent work by the way. I wish I knew how to code!

Edit: I should add this here for high visibility.

If you have an MCAT 2015 score but would like to use this app, check out my conversion tables to tell you your percentile equivalents for the old sections.
 
Last edited:
OP
paul411

paul411

ANES
7+ Year Member
May 27, 2010
1,616
33
Florida
Status
Resident [Any Field]
It's the weighted avg of GPA and MCAT. And within the GPA, the science GPA (0.65) is weighted more than the cumulative GPA (0.35)
 
About the Ads

TheOtterpop

Not an otter pun!
Jun 8, 2015
62
45
Colorado
Status
Medical Student
Looks pretty nifty! What about a "sort by chances" button or something? Markers for public/private schools?
 
OP
paul411

paul411

ANES
7+ Year Member
May 27, 2010
1,616
33
Florida
Status
Resident [Any Field]
I'm working on the "sort by" feature. I like the public/private idea, I'll add that too. Although, I think cost of attendance is a better data point because public/private is basically a surrogate for cost.
 

inky

2+ Year Member
Jun 17, 2015
156
140
Status
MD/PhD Student
Wow, this looks amazing.

Maybe you could add a section for graduate GPA? I know that isn't weighted as highly as anything else but for some reinvention people it can be important.
 

GrapesofRath

2+ Year Member
May 5, 2015
5,320
3,803
Status
Non-Student
Interesting work for sure. Must have taken alot of effort. One thing I'll say is that the stats you have for each school are clearly not from MSAR. Some of them appear to be from school's websites. Either way they appear rather outdated. For example no way Eastern Virginia's median stats now days are 3.44/30(which is from their site). I think it's honestly closer to 3.65/31-32. Like wise there are many other schools where the data is significantly off; ie no way in hell is NYU a 3.75/33. It's closer to a 3.9/37. I'm not sure how that plays into your algorithm but it's something to consider.
 

Lawper

Cat-box cycle
Gold Donor
5+ Year Member
SDN Ambassador
Jun 17, 2014
38,132
111,525
space chat
Interesting work for sure. Must have taken alot of effort. One thing I'll say is that the stats you have for each school are clearly not from MSAR. Some of them appear to be from school's websites. Either way they appear rather outdated. For example no way Eastern Virginia's median stats now days are 3.44/30(which is from their site). I think it's honestly closer to 3.65/31-32. Like wise there are many other schools where the data is significantly off; ie no way in hell is NYU a 3.75/33. It's closer to a 3.9/37. I'm not sure how that plays into your algorithm but it's something to consider.
MSAR is copyrighted.

All the data is sourced from school websites and publicly available data.
 

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
I'm working on the "sort by" feature. I like the public/private idea, I'll add that too. Although, I think cost of attendance is a better data point because public/private is basically a surrogate for cost.
The cost/debt numbers may be extremely misleading as the SES distribution is likely very different among schools. The adcoms on here advise against putting much weight on average debt numbers when compiling a list to apply to, and say go ahead and apply and (if accepted) wait for your personal financial package.

One other tiny nitpick: When you mouse over academic index it is not immediately clear that it is telling you how much above or below the number you are. Especially for the academic index numbers mousing over "500" and seeing a popup say "your academic index is 400" sends exactly the wrong message! Could make it "your academic index is 400 points higher". For the lower cases it is more clear thanks to the "-" in front of the number, but still would be nice to have "is X points lower"

Sorting would be awesome! I suspect this will become sticky-worthy once it's all tuned up

One last thing: how do you plan to keep the numbers up-to-date? You could add contact info or a submissions box where people can send you updated stats and a (free non-MSAR) source
 

GrapesofRath

2+ Year Member
May 5, 2015
5,320
3,803
Status
Non-Student
MSAR is copyrighted.
Well the question here is how data that is clearly off affects the algorithm, not issues regarding copyright. I never said anything about using MSAR's data in an unauthorized manner, rather consider how data that clearly isn't up to date affects the algorithm and its calculations, if it does at all.
 

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
Well the question here is how data that is clearly off affects the algorithm, not issues regarding copyright. I never said anything about using MSAR's data in an unauthorized manner, rather consider how data that clearly isn't up to date affects the algorithm and its calculations, if it does at all.
Do schools not all have their mcat and GPA on their sites? Maybe this is just an issue of needing some updating
 

md-2020

The Immaculate Catch
2+ Year Member
Jun 29, 2015
2,298
3,009
Status
Medical Student
You should add a filter by race.


Not trying to stir up more URM debates, btw. It's just more accurate to factor this in.

Edit: Also, you mislabeled UPenn and Penn State. You currently have 2 Penn St.'s, one of which is clearly Perelman's stats
 
About the Ads

GrapesofRath

2+ Year Member
May 5, 2015
5,320
3,803
Status
Non-Student
Do schools not all have their mcat and GPA on their sites? Maybe this is just an issue of needing some updating
Alot of them are clearly outdated. And the reason I bring it up is their outdated stats are often a good bit different than what they currently are. For example if someone sees Eastern Virginia's average GPA of 3.44 as listed here or their site and think they got a good shot with a 3.4 when in reality the median GPA is closer to a 3.7, that's rather misleading. There are a number of other examples like that you could find here.
 
  • Like
Reactions: efle

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
You should add a filter by race.


Not trying to stir up more URM debates, btw. It's just more accurate to factor this in.

Edit: Also, you mislabeled UPenn and Penn State. You currently have 2 Penn St.'s, one of which is clearly Perelman's stats
I don't think any schools release matriculant data per race...
 
OP
paul411

paul411

ANES
7+ Year Member
May 27, 2010
1,616
33
Florida
Status
Resident [Any Field]
Alot of them are clearly outdated. And the reason I bring it up is their outdated stats are often a good bit different than what they currently are. For example if someone sees Eastern Virginia's average GPA of 3.44 as listed here or their site and think they got a good shot with a 3.4 when in reality the median GPA is closer to a 3.7, that's rather misleading. There are a number of other examples like that you could find here.

"In recent years, students matriculating at EVMS have had a mean GPA of 3.44 and a mean MCAT of 30" From the EVMS website. Where are you getting the 3.7?
 

nhnative

2+ Year Member
May 6, 2015
87
52
Status
Medical Student
This is a cool idea, nice work! I think UNE (osteopathic) is off as well, their MCAT definitely isn't 32.... from their website

Grade Point Average (GPA) has been above 3.5 for several years (with science GPA at 3.45 for several years). The Committee on Admissions prefers that applicants have scored at least a 3.4 overall (catalog-published minimum is 2.7). The average for the Medical College Admissions Test (MCAT) has been 27 for several years.
 

musicalfeet

5+ Year Member
Feb 1, 2013
1,047
738
Status
Medical Student (Accepted)
I think it does relatively well in estimating your chances? So far all my II came from schools labeled with "Good chance", but also a rejection from a "high chance" school. I find that the list it generated is pretty close actually to the one on prospectivedoctor.com, which so far has been pretty accurate in predicting interview/hold/rejections for me. Still early in the cycle to say anything definitive though.

According to this though, I picked a pretty good list of schools, with the majority of them being "good chance". Ha we'll see how that plays out!
 

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
There's a spreadsheet online that somehow incorporated the race factor, and I found the results to be quite reasonable. Here it is: http://www.prospectivedoctor.com/chance-predictor/

It's probably the most comprehensive/analytical chance/LizzyM predictor I've come across.
Hah, this doesn't strike me as particularly reasonable :p

Cool site though. Now how are they getting all this data without violating MSAR or US News copyright...
 
  • Like
Reactions: md-2020
About the Ads

gyngyn

Staff member
Administrator
Moderator
Lifetime Donor
7+ Year Member
Nov 4, 2011
24,219
40,485
Status
Attending Physician
There's a spreadsheet online that somehow incorporated the race factor, and I found the results to be quite reasonable. Here it is: http://www.prospectivedoctor.com/chance-predictor/

It's probably the most comprehensive/analytical chance/LizzyM predictor I've come across.
I plugged in some real stats of actual students and found it to be way off (esp. on the ethnicity issue).
At my school it would have excluded most of our UIM students.
The only choice for most Latinos in this schema is "Non-white Hispanic or Latino!"
This would be a very small subset of the Latino applicant pool.

They also failed to account for IS biases and mission.
 
  • Like
Reactions: md-2020

musicalfeet

5+ Year Member
Feb 1, 2013
1,047
738
Status
Medical Student (Accepted)
Hah, this doesn't strike me as particularly reasonable :p

Cool site though. Now how are they getting all this data without violating MSAR or US News copyright...
So far that's the one that's predicted most accurately where I'd get II (as well as holds/rejections!). Although I'm a pretty skewed student in terms of GPA (low ugpa...so I brought my cgpa to average by my kickass post-bac gpa..which in turns mean my sgpa is really high too) which is why the LizzyM (because it uses cgpa only) may not be as consistent for me?
 

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
So far that's the one that's predicted most accurately where I'd get II (as well as holds/rejections!). Although I'm a pretty skewed student in terms of GPA (low ugpa...so I brought my cgpa to average by my kickass post-bac gpa..which in turns mean my sgpa is really high too) which is why the LizzyM (because it uses cgpa only) may not be as consistent for me?
You are correct! LizzyM is a barebones approach that doesn't factor in upwards trends, SMPs etc
 

md-2020

The Immaculate Catch
2+ Year Member
Jun 29, 2015
2,298
3,009
Status
Medical Student
Hah, this doesn't strike me as particularly reasonable :p
Well I mean if you put in 4.0/45 statistically you would be an overshoot at HMS, which averages 3.9/37. A spreadsheet can only go off numbers, which will admittedly be inherently flawed. It will always be an approx.
 

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
Well I mean if you put in 4.0/45 statistically you would be an overshoot at HMS, which averages 3.9/37. A spreadsheet can only go off numbers, which will admittedly be inherently flawed. It will always be an approx.
They should just put "Don't get your hopes up" for H S JH regardless of stats you put lol
 

WedgeDawg

not actually a dog
Staff member
Administrator
7+ Year Member
Mar 22, 2012
7,644
12,423
Status
Resident [Any Field]
OP
paul411

paul411

ANES
7+ Year Member
May 27, 2010
1,616
33
Florida
Status
Resident [Any Field]

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
I just emailed the admissions coordinator at EVMS and she confirmed that the stats on their website (3.44/30) are the most current. I'm not sure where the MSAR data is from.
The MSAR data is reported from the office of admissions as well. This could be a difference in accepte vs matriculant data, as you'd expect the curve of matriculants to be shifted slightly lower than the curve for acceptances
 
OP
paul411

paul411

ANES
7+ Year Member
May 27, 2010
1,616
33
Florida
Status
Resident [Any Field]
The MSAR data is reported from the office of admissions as well. This could be a difference in accepte vs matriculant data, as you'd expect the curve of matriculants to be shifted slightly lower than the curve for acceptances
Here's definitive matriculant data for EVMS 2014 MD entering class: https://www.evms.edu/media/facts&figures-2014-2015-2014.pdf (page 8.). Average GPA 3.56 Average MCAT 31. I don't actually have access to the MSAR. Is this close to the MSAR data?
 

Dr.TonySoprano

5+ Year Member
Jun 8, 2014
524
364
Status
Medical Student (Accepted)
I've created a web app (from the data in the SDN-used created spreadsheet) to help people applying to med schools pick their schools. The app allows you to input your stats (GPA, MCAT) and it calculates your chances of getting into all the medical schools based on the academic stats. It also information about in-state vs. out-of-state numbers and costs. All the data is sourced from school websites and publicly available data.

You can access it here: Medical School Data App at http://paul411.github.io

It's in its infancy. I'll add more data and options (e.g. comparing schools) in the future. Let me know if you think of any suggestions and/or improvements.
you should try using the mcat 2015 scores too in your system, but that would be even more difficult to predict
 

GrapesofRath

2+ Year Member
May 5, 2015
5,320
3,803
Status
Non-Student
Here's definitive matriculant data for EVMS 2014 MD entering class: https://www.evms.edu/media/facts&figures-2014-2015-2014.pdf (page 8.). Average GPA 3.56 Average MCAT 31. I don't actually have access to the MSAR. Is this close to the MSAR data?
One thing to note is that MSAR we are referring to is median data not mean data. Median data is more accurate IMO. I don't have MSAR in front of me but I believe Eastern Virginia's MSAR data is around something like median GPA: 3.67 and median MCAT as a 31 or 32.

Note I'll add alot of the data listed is lower than what MSAR will tell you and Eastern Virginia is hardly the most glaring example(NYU is at 3.87/37 per MSAR not 3.75/33 huge difference). I'm not sure how it affects your model or if it's really relevant to what you are trying to get at, I figure I'd just let you know.
 
  • Like
Reactions: gyngyn

gyngyn

Staff member
Administrator
Moderator
Lifetime Donor
7+ Year Member
Nov 4, 2011
24,219
40,485
Status
Attending Physician
The MSAR data is reported from the office of admissions as well. This could be a difference in accepte vs matriculant data, as you'd expect the curve of matriculants to be shifted slightly lower than the curve for acceptances
The AAMC calculates the matriculant data from the source, we don't report it.
US Snooze uses our reported "data."
 
  • Like
Reactions: efle

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student

gyngyn

Staff member
Administrator
Moderator
Lifetime Donor
7+ Year Member
Nov 4, 2011
24,219
40,485
Status
Attending Physician
TIL! MSAR is accepted student data ranges rather than matriculant though right?
The MSAR uses their own AAMC primary data for the stats in the MSAR (not our wishful thinking "data").
Yes, the ranges are for accepted students.
Matriculants can actually be a bit higher or lower, but it's pretty close. US Snooze is reported matriculants data (from the school).

What would really be interesting is the median scores of waitlisted candidates! I'll bet ours is higher than accepted students.
 
  • Like
Reactions: WedgeDawg and efle

efle

not an elf
5+ Year Member
Apr 6, 2014
11,693
15,514
Status
Medical Student
The MSAR uses their own AAMC primary data for the stats in the MSAR (not our wishful thinking "data").
Yes, the ranges are for accepted students.
Matriculants can actually be a bit higher or lower, but it's pretty close. US Snooze is reported matriculants data (from the school).

What would really be interesting is the median scores of waitlisted candidates! I'll bet ours is higher than accepted students.
What on earth would make for a higher matriculant median than accepted median? Aren't the right end of the curve much more likely than the left end to have alternative attractive options?
 

GrapesofRath

2+ Year Member
May 5, 2015
5,320
3,803
Status
Non-Student
I would expect the biggest differences in matriculant vs accepted data to be in the 10th-90th percentile data honestly. When you are comparing a sample of say 250 accepted students to only 110 matriculated students, that changes things more along those lines. There are just too many 3.75/32 types at lower tiers, 3.8/35's at medium tiers and 3.85/37 types at top tiers to really change the mean/median that much when comparing accepted to matriculated median and means much.

I certainly would expect the 90th percentile data of matriculated students to be lower than accepted, particularly when talking about lower and medium tier schools. To a lesser extent, the 10th percentile data might be slightly lower for matriculated students than accepted students. I could actually see the biggest gap for this at state schools which have more incentive to interview and accept those higher stat applicants who probably won't attend than your typical lower tiers.

My guess is that the whole waitlisted candidates having higher stats than accepted is school specific and a trend for gyngyn's school not the majority of other schools. I'm skeptical that it has anything to do with yield protection; unless it's blatantly obvious a top candidate has absolutely zero interest in attending through the interview, I doubt any school would take the time to interview someone they think has very little chance of attending unless they really really like the candidate.
 

gyngyn

Staff member
Administrator
Moderator
Lifetime Donor
7+ Year Member
Nov 4, 2011
24,219
40,485
Status
Attending Physician
What on earth would make for a higher matriculant median than accepted median? Aren't the right end of the curve much more likely than the left end to have alternative attractive options?
I would expect the biggest differences in matriculant vs accepted data to be in the 10th-90th percentile data honestly. When you are comparing a sample of say 250 accepted students to only 110 matriculated students, that changes things more along those lines. There are just too many 3.75/32 types at lower tiers, 3.8/35's at medium tiers and 3.85/37 types at top tiers to really change the mean/median that much when comparing accepted to matriculated median and means much.

I certainly would expect the 90th percentile data of matriculated students to be lower than accepted, particularly when talking about lower and medium tier schools. To a lesser extent, the 10th percentile data might be slightly lower for matriculated students than accepted students. I could actually see the biggest gap for this at state schools which have more incentive to interview and accept those higher stat applicants who probably won't attend than your typical lower tiers.

My guess is that the whole waitlisted candidates having higher stats than accepted is school specific and a trend for gyngyn's school not the majority of other schools. I'm skeptical that it has anything to do with yield protection; unless it's blatantly obvious a top candidate has absolutely zero interest in attending through the interview, I doubt any school would take the time to interview someone they think has very little chance of attending unless they really really like the candidate.
In CA we have tons of high stats applicants. The waitlist is full of high stats.
The lower stats applicants with qualities strong enough to distinguish themselves in this competitive environment are the ones that everyone wants. They are the ones with multiple accepts and many options (like recruitment scholarships).
 
  • Like
Reactions: GrapesofRath
About the Ads