A media buzz has started last weekend with the public launch of WolframAlpha. If you haven't heard about it yet, WolframAlpha is a new web search engine. What makes it the focus of media attention is that it is the brainchild of British physicist and mathematician Stephen Wolfram - the man behind Mathematica.
Search engines come and go but for several years now Google has dominated web searches in Australia and the rest of the world. WolframAlpha's claim to fame is that it can answer questions that Google can't: those that require higher levels of natural language processing and integration of data sets. This is best explained using an example: Say you want to compare the financial performances of IBM with those of SAP. Using Google or Yahoo! (for example) you can build mini-profiles of each company using stock market and SEC filings. WolframAlpha goes a step further by throwing more data sets into the mix. Here is the basic IBM versus SAP result from WolframAlpha:
WolframAlpha-SAP IBM Search Result PDF
WolframAlpha enriches the search engine findings in two important ways. Firstly it mines more 'specialist' data sets and integrates them. In this example the potential data sets used included:
Figure 1: WolframAlpha Sources
In this example, this means that fundamental analysis can include:
- Common ratios
- Balance Sheet
- Income Statement
- Cash Flow Statement.
So the data set available to the user is richer (more data = richer):
Figure 2: Fundamentals Analysis
Secondly it leverages functionality from the Mathematica analytics package. So the data set available to the user is more powerful (more statistical tools = powerful). This means that you can slice and dice the data in more flexible ways:
Figure 3: Volatility Analysis Figure 4: Portfolio Optimization
There is much more to WolframAlpha and it is still early days.
I expect it will be a while before the full power of WolframAlpha is apparent.
The example I have used is typical of the analytics I make use of but I suspect that the impact of WolframAlpha
will be most apparent to 'simpler' questions from the general public.
Questions like:
The actual service still feels like a beta and response times can be frustrating but I expect that these are teething
problems that will be resolved. Competing against Google must be an incredible challenge.
It's like trying to beat Microsoft. Maybe WolframAlpha is the future Apple/Linux of the search world?
My take on WolframAlpha is that it is potentially a great new addition to web-based analytics. Whether-or-not it is
the next big thing will depend on how quickly it can evolve into a polished end-user service. The potential is clearly there
as the founder has a solid track record in building Mathematica into a leading product.The next year of search engines just got a lot more interesting.
