How do you get software to be able to access stock market data?

>>1356553
Lmao that's cute Photoshop - the original had a kitten.

File: 1444419757765.jpg (25KB, 403x403px) Image search: [Google]

25KB, 403x403px

> "play with AI a little"

Let me get this straight. You know how to codify and modify artificial intelligence. You have access to and know how to fine tune a warehouse modeling engine, a predictive analytics engine, business rules engine, and an underlying engine with slowly changing dimensions ... and you decide to take this resource and engage the 4chan community in order to gain their perspective on how to successfully invest in the stock market, because you'd rather not study it yourself.

Yea, you're not full of shit at all.

>>1356939

I tend to agree with this gentlemen's opinion.

https://www.quantopian.com/

Enjoy, OP - literally everything you want to ask can be found there

>>1356939
What's with asses like these every time you make a post like this?
I'm not wanting to make some super-optimal business AI, I'm just wanting to screw around with different learning algorithms, and all I need to know is how raw stock information is ported to AIs.

i just use excel and a query function:

http://www.msn.com/en-us/money/quoteslookup?symbol=vdc

im sure there are other ways

>>1357046
you're trying to describe a trading algo. just go buy one for $250k and be done with it, go get dat dosh boyo

>>1357067
Not a trading algorithm, I just want to make a generic learning AI and test it by putting it to imaginary trade (doesn't buy stocks, instead simulates buying stocks).

>>1357073
so, instead of connecting the bot to a legit account, you fake a connection to a real account. literally no difference you fucking idiot.

youre asking about either: a program someone built over months; a program which costs hundreds of thousands of dollars

>>1357076
I think the point is so they can get a fun programming learning exercise out of it. Why you gotta be grumpy about an innocent question qt?

>>1357076
>literally no difference
Picture it this way.
The AI doing real gambling will bet real money on two combatants in the arena, and if it's a good AI it'll make some income by identifying the better combatant, while if it's a bad AI it will cost money to run.

An AI doing imaginary gambling will bet imaginary sums of money on the combatants, and that imaginary money is treated as real by the AI, but the money doesn't actually exist and any income or loss is also imaginary.

Also, simple learning algorithms can be made, as I said I'm not looking to spend months on a good business AI; just looking for practice.
I'd start with something really simple like an algorithm that looks at a thousand company stocks, and tries drawing relationships saying things like "If company 193's stocks go up, company 925's will go down".
Extremely rudimentary and inefficient, but incredibly easy to program and there's no harm in an AI that only pretends to trade; in much the same way a child playing monopoly isn't a disaster.

>>1356553
>simple programming wise
As a programmer I want to say you are fucking stupid.

>>1356553
use fuzzy computing make your bot aggressive
http://www.wired.com/2016/06/ai-fighter-pilot-beats-human-no-need-panic-really/