There's a website called Typealyzer that claims to analyze text and return the likely MBTI type of the writer. The Myers-Briggs Type Indicator is commonly called a personality test, which makes it sound a lot broader in scope than it is. Rather than encompassing a person's entire personality, it looks at how you take in information and how you make decisions, which could be extrapolated to quite a lot, but not the entirety.

Bear in mind that I didn't make Typealyzer, so I don't know how it works at the back end. I have no idea whether its results are accurate or meaningful in any way, so this is all very unscientific and just for fun, though it may be useful or interesting to you in the same vague, subjective way it was for me.

It would have taken way too much work to do all of this manually (some of these users have over a thousand writeups) so how is this best done? Perl, of course. One script takes an input file containing the list of MBTI types and users from the other writeup in this node. It creates another file containing the URLs to all of those users' writeups. I added myself to the ISTP section for the first run, but then I found out I was doing something critically wrong, and I forgot to add myself for the second run. And now all the data is finished and I'm tired of dealing with it, so I get to sit this one out.

Once this list of URLs is ready, another script reads it. The list has lines indicating the currently expected MBTI type and the user whose writeups we're looking through. When the script encounters a line containing an MBTI type, the expected type is updated; same for the current user. These are inserted into the final file. This could have been done with some fancy parsing of the URL or so, but a good programmer is a lazy programmer. When this script finds a line that it determines to be a writeup URL, it fetches that webpage, isolates the actual writeup content, takes the MD5 hash of that content, stores a copy of the content to webspace with the hash as the filename, submits the URL to the isolated content to Typealyzer, and extracts the MBTI type from the response. This MBTI type that Typealyzer reports for the content forms part of a new line of a CSV file, along with the hash-filename, the type the user is supposed to be, and the original URL to the writeup on E2.

These scripts can be combined into a single script to where, theoretically, you can just copy the list of types and users from the other writeup in this node, feed it to the script, and end up with the final CSV file. In practice, a URL fetch fails every few hundred or thousand writeups, or something else goes wrong, and the script has to be restarted. It's much easier to fix the intermediate file containing the list of writeup URLs, then restart the second-stage script.

The final list of results will be sent to the owner of Typealyzer and can also be used for fun and profit. For example:

Number of writeups by the user's expected type:

ISTJ: 2,530
INFJ: 1,241
INTJ: 4,836
ISTP: 232
ISFP: 63
INFP: 4,378
INTP: 10,257
ESTP: 249
ESFP: 166
ENFP: 834
ENTP: 3,990
ESTJ: 52
ESFJ: 254
ENFJ: 1,805
ENTJ: 629

Number of writeups by the type Typealyzer reported:

ISTJ: 845
ISFJ: 776
INFJ: 417
INTJ: 1,750
ISTP: 2,377
ISFP: 2,136
INFP: 2,531
INTP: 6,367
ESTP: 2,392
ESFP: 5,432
ENFP: 1,278
ENTP: 1,321
ESTJ: 613
ESFJ: 2,280
ENFJ: 426
ENTJ: 497

If we compare writeups by expected type with writeups by reported type, we can see that the reported types are somewhat more evenly distributed, though the propensity of the INTP to ramble is still well represented.

Of 31,518 writeups, there are 3,501 (11.1%) for which the expected type matches the result. By type:

ISTJ: 52
INFJ: 25
INTJ: 334
ISTP: 23
INFP: 302
INTP: 2,304
ESTP: 31
ESFP: 38
ENFP: 24
ENTP: 297
ESFJ: 33
ENFJ: 20
ENTJ: 10

Top ten users by number of writeups where the expected type matches the result:

463: The Custodian
359: Pseudo_Intellectual
221: ModernAngel
185: sekicho
149: Uberfetus
148: mauler
132: enth
114: JerboaKolinowski
92: Evil Catullus
85: liveforever

mauler pointed out that some people made this list just by having a ton of writeups period, so:

Top ten users by percentage of their writeups for which the expected type matches the result:

66.67: Aphrodite
54.81: BurningTongues
53.57: NinjaPenguin
48.84: WolfDaddy
42.70: exceptinsects
41.32: grundoon
36.90: Jack
36.87: LaylaLeigh
36.70: momomom
33.85: mr100percent

Typealyzer result for this writeup: ENTJ