₦airaland Forum

ishmael: I'm happy u guys are back. This is a nice thread i must say. The truth is that many computational courses like maths, physics, stats, ops research, etc are heavily dependent on computers today to carry out computations efficiently, therefore there is need for them to have some knowledge of computing.

Sure. But I might use a toaster to make my breakfast. That doesn't mean that I am closely related to a chef, or something.

Everyone needs to know some computing nowadays. But that doesn't mean that they are defined by it, or that is their career.

The boundary between pure and applied can be somewhat unclear at times. And then you have famous mathematicians like Terence Tao who have done excellent work in both areas.

Definitely in the 40s, 50s, 60s, you'd have guys who were pure mathematicians making huge contributions to more applied fields.

The early statisticians were all mostly mathematicians, right?

Same for the early computer scientists.

The Nash equilibrium guy had his training in pure math.

I don't know if this is true in physics as well, though (and suspect that it is not).

Sure. Cryptography, RSA (used whenever you buy stuff online or visit a https:// website) is applied number theory. But I view that primarily as applying existing theory to solve concrete problems (though, this is by no means easy or trivial.

I guess it depends on what you define a "pure mathematician" to be. But most statisticians are motivated by some sort of real applications, real-world data, or existing procedure that they want to analyze or describe the properties of.

Pure mathematicians, I guess my conception of them is people who work on relatively obscure problems in math that have absolutely no immediate application or use.

Statistics is a bit different...even a theoretical statistics paper which for example proves that "procedure X is asymptotically consistent under the following conditions", well, that is kind of immediately useful, since it means that procedure X behaves well as you give it more data. Which of course is useful to anyone who is using procedure X to solve a real problem.

Statistics isn't really a branch of pure math.

Probability is, yes. Statistics is more applied math.

Perhaps we should make a list of core topics that one sees in an undergraduate statistics major.

How many of them are "computer sciency" in nature?

They are three distinct fields, but obviously have areas of overlap.

Physics and statistics: MCMC techniques/algorithms (and probably lots of others)
CS and statistics: machine learning, randomized algorithms
Physics and CS: quantum computing, probably a few others

Statistics, well it is definitely a subset of math in general. Tools like linear algebra, optimization, probability theory (essentially measure theory), calculus, etc all come into play to justify statistical methodology (e.g., why is the central limit theorem true? when is the Poisson approximation for the binomial distribution valid? etc).

Statisticians don't really care much about the algorithmic aspects? E.g., convergence, running time, etc. Computer scientists, especially theoretical computer scientists compare about these issues.

I dunno, just read through like a leading statistics journal. The papers there aren't very similar to what one would read in a CS journal.

Nor are they similar to most of what will read in a physics journal.

Which brings me back to my main point...these are three distinct fields. It doesn't make sense to two of them especially go together.

Yes, statistical work requires the use of computers nowadays. So does a lot of mechanical engineering, physics, chemical engineering work.

I just think that if you are a writer, part of your goal should be as clear as possible to your audience. Using reasonable punctuation, indentation and proper English isn't "forming." It simply makes it easier for your readers to understand you.

The statisticians I know do pretty simple stuff on the computer. PCA, regression/linear models, etc. Things that are prepackaged in statistical software.

They are no more programmers than a mechanical engineer who builds his models in matlab, or a chemical engineer who builds his models in mathematica is a programmer.

Being a consumer of software and writing the occasional small script doesn't make you a computer scientist.

freecocoa: Him talk say una too dey cuss person for here because of small mistake in the grammatical aspect,say e no good instead make una correct the person with love because why?no be we papa dem get english but people here dey do like say dem sabi spoke English pass queen Eliza sef,say the thing dey make am para wella.

Something like that,i no read pass 3 lines

lol

alj harem: thanks you all, I am alive and I am not part of BOKO HARAM

good!

Can someone translate the op's post into English for me. I cannot read what he wrote

Agiliti: you are the only cockroach in my cupboard
language = local

Interesting.

Nice thread.

In nigeria perhaps. But elsewhere, this is not the case

There is a very generic between statistics and computing, in that most statistics is done on a computer. But statisticians aren't familiar with the same concepts one would expect from a CS graduate.

Nothing wrong with weed. But, for me it makes me dumber than usual. So I really can't afford to smoke it.

Not to mention drug screenings at jobs, etc

Better question is, why are women so mean to each other? Always trying to pull each other down..

And not too many develop new statistical algorithms. Just reuse known stuff. E.g., run PCA on such-and-such dataset.

Haha not the ones I know.

They are pretty meh programmers.

Haven't you noticed how ugly their main data analysis tool (R) is?

I'm not sold on statisticians naturally having some affinity for programming

Not true at all. And I speak from experience.

Most of us would do the same if we were in their shoes.

Jesus was a black man from Oyo State, in Nigeria. In fact, we are distant relations.

Ok. Maybe in Nigeria the majors are similar. But in my experience, in the US they are not.

Most stat students I know aren't very good at programming, for example. Don't know much about algorithms and data structures. Etc.

Not to say that physicists know these things either.

Basically, in my experience, they are three pretty separate and distinct courses of study (at least in the US)

A coach in one of the teams admitted that Nigeria lacked world class athletes and that even if more billions of money were approved the money would have still gone down the drain as “you cannot perform magic over night in sports, you build over a period of time.”

A country of 170+ million black people lacks world-class athletes

Wtf?

Lacking the athletes is not the issue. Finding them and nurturing them is.

What programming courses do stats students take in the UK and Nigeria?

What are the mandatory math and stat courses for CS students, aside from those pretty much every engineering student must take?

Katsumoto: Obafemi Awolowo by a thousand miles.

+1

database? statistics? like an sql database?

that isn't how it works in the US at most schools. You start taking computing courses as a CS major pretty early...freshman year.

Statistics isn't mandatory either for CS students, at least at the universities I have experience with.

Pretty much no core CS or programming classes are required for statisticians either.

ishmael: Just say u don't know. Do a little research and find out the courses statistics students and computer science students offer. Then find out if there is a correlation.

I don't know the mandatory courses for those degrees in Nigeria, but I do know what they look like in say the US. And there is no more overlap between CS and statistics in those places than there is between CS and physics.

[quote author=Ricky_Ross]INDIANS ARE AS CORRUPT AS NIGERIANS.[/quote]But their country is growing more rapidly than Nigeria.

So, what conclusion should we draw from this?