Wolfram|Alpha is a very interesting new search and analysis technology that was officially launched in the last week or so. Calling itself a "computational knowledge engine", it could revolutionize what you think about search engines like Google, and how you access information on the Internet.... someday. To say it's very ambitious (and it does), is something of an understatement, but it's already an amazingly cool tool, and shows some great promise for any area that deals with data collection and analysis, or scientific research.
Alpha is the brainchild of Stephen Wolfram who developed one of the most advanced applications ever based on Artificial Intelligence programming techniques, Mathematica. Mathematica allows users to input extremely complex math equations including calculus, differential equations, infinite series, etc., and then uses symbolic analysis and rules to solve the equations and/or plot the analysis. Many equations like this require tricks/intelligence to solve, rather than algorithmic approaches, and Mathematica has been a boon to those in the fields of physics, chemistry, and engineering for nearly three decades.
In my college days I actually studied Artificial Intelligence, and have long been an admirer of Mathematica. So, I was very interested last year when I heard about Wolfram|Alpha, and even more so when it was finally launched to much fanfare recently. Wolfram|Alpha allows you to enter a search phrase, and instead of returning a bunch of existing pages, it seeks out data and assembles a new page or analysis for you on the fly. Depending on what it decides you're looking for, it may create a chart, a map, a variation of a graphic, or tabular data. In their own words:
"Wolfram|Alpha's long-term goal is to make all systematic knowledge immediately computable and accessible to everyone. We aim to collect and curate all objective data; implement every known model, method, and algorithm; and make it possible to compute whatever can be computed about anything. Our goal is to build on the achievements of science and other systematizations of knowledge to provide a single source that can be relied on by everyone for definitive answers to factual queries."
How's that for an ambitious goal? If anyone else other than Stephen Wolfram was behind it, it would be laughable. But based on what he's accomplished with Mathematica over a few decades, and applying the same sort of approach, I think anything is possible. Just don't expect the world today. It may be several years before Wolfram|Alpha is truly ready to amaze. [When it's a little more mature, will it be called Beta? Just asking...]
One of the most interesting application areas for it already is financial data analysis. While there are some interesting sample queries on the site to show what it can do, I decided to try a few of my own. An easy data collection and analysis report that can be generated is comparing the current market caps of two stocks (in this case Toyota and Ford, a sample from their site). As you can see, we've got a simple math equation in the search form, and a historical chart is created on the fly (it just assumes that we need a historical chart along with the current value, so some decisions are being made for you about what you need most, and if you need data prior to 1989, well, you're on your own):
Another simple query I asked for is to see the current yield curve on the US Treasury. There had to be some interesting data it came up with, and good charts that would be assembled on the fly. Based on the simple query, here's what I got (simple, straightforward, and informative):
But just to show that there's a lot of evolution and development ahead of them, Wolfram|Alpha didn't have such good luck with a Corporate bond yield curve. I tried several ways of asking the question, but it couldn't figure out what I was asking for, and the suggestions were only partially helpful, if not completely irrelevant:
I guess when you're new and trying to make a good impression, but you get thrown a curve you can't handle, it's a natural response to go with your strength. Nevertheless, I don't need the encouragement to explore curves and surfaces, or weather and meteorology when I'm trying to get a graph of a yield curve.
I also found other immediate limitations in that there was no knowledge of ETF symbols, just individual stocks. I asked for GLD as a proxy for the price of gold, and got Gladstone (traded on the London exchange). I then asked for information about gold specifically and got the atomic number and its place in the periodic table.
As an exercise for the reader, check out the comparison charts of multiple stocks, like GOOG, MSFT and AAPL. Not as detailed as you'd get on Yahoo! Finance, but stylish and informative none the less. Clearly Wolfram|Alpha is not a one trick pony, but it's a small finite number of trick ponies when it endevours to handle a seemingly infinite number of ponies. It's pretty slick now, will probably be incredibly impressive in about 3-5 years, and fully baked in about 2050. I'll let you know when it's ready for Beta.


Recent Comments