Putting Wolfram|Alpha to the test against Google
Every few months there seems to be a new company that releases a new product to the market that is a going to change the whole world with a simple few lines of code maybe this is the beginning of that process? Google the world’s largest search engine has an academic background being born from two students at Stanford University before starting Google in 1998, it is also the source of a very recent public failure within its European market.
One of the founders of Google Sergey Brin worked at Wolfram Research back in 1993, well before even Netscape allowed the internet to be accessible to everyone. Sergey was involved in the original code analysis and extraction tool for Mathematica which is a paradox because it seems Wolfram|Alpha which is built on the Mathematica language, may be more advanced that Google and is now live.
While Wolfram|Alpha is not a new product, with social networks buzzing with news, screenshots and reviews for months it is finally live and they are taking feedback on how people are finding the initial results. This welcome interaction to help deliver a product, shows that this is not a static product but still may have some more refinements to its results. Afterall, Google is still refining their results after 11 years.
You might wonder what is the relevance to covering a new search engine when people are content with googling everything these days? Well, Wolfram|Alpha is a computational-based knowledge engine, so is less prone to spam, manipulation of results or censorship of data/dates.
As this is a new product Wolfram|Alpha has a suggestion box on the right side of the results suggesting a new things to try so you can test the power of its knowledge.
- Enter any date (3 April 1975)
- Enter any town (Brisbane)
- Enter any two stocks (Telstra & ATT)
- Enter any calculation ($250 + 10%)
Using the suggested things to try, I found the results were amazing with sample stock search gathering data from ASX and NYSE and combining the data into a spreadsheet so you can visually see if Telstra’s drop in share price is a sign of the international market or it is localised. Also the Wolfram|Alpha share search is useful for reviewing two companies from different markets quickly and making performance comparisons. These data-focused results are perfect for business people and journalists who are after quick access to current data for reports.
To test how knowledge based search would compare against commercial search, I ran a sample search with three keywords to determine which search engine would provide more detail since this a research/information query, you can see the side by side screen shots in the images at the bottom of this article.
Test 1: Wolfram|Alpha vs. Google ISS
Three suggested results: NASA, 2 x Wikipedia, and several for security company ISS within the top results.
Current map position on globe, current orbital information, launch date, sky position from Brisbane. It also shows all the data in metric terms with other detail such as if it is currently visible in the sky above Brisbane, and explains in abbreviations of terms such as rev/day. Each section also has a more link that expands to show dozens more details such as: velocity, altitude, orbital period, sun angle, inclination, mean motion, launch vehicle, NORAD #, etc.
The Wolfram|Alpha team understand that people are comfortable with Wikipedia so they offer a link to the official Wikipedia page for that particular topic.
Test 2: Wolfram|Alpha vs. Google Brisbane to Los Angeles
Both paid/organic results for commercial services, including flight services, between the two locations but no travel information.
Distance information, flight paths, local times and population statistics along with related links to Wikipedia pages. The search results also show any assumption that the search engine made such as you were talking about LA, USA not LA, Chile or LA, Philippines.
Test 3: Wolfram|Alpha vs. Google chef average wage
Some relevant results for salary information, but that is also mixed in with commercial results, books and blogs.
2007 estimates, historical graph of pay and employment summary. However, the service lets only provides US-specific information for wage/salary related questions at this time.
The Wolfram|Alpha search engine offers some amazing potential improvements for academic, research and informational enquires. Featured at the bottom of the search results is a source reference link that shows where all the data that was presented was sourced from, which allows for more transparency in results. Another cool feature is the ability to export the search results as a PDF document for reviewing or printing later.
Google will remain the dominate player for many years, Google also developed its product and userbase by equipping academics as its early adopters with volumes of data that was much more relevant than the other search engines. So the question is has Google sacrificed its academic background providing the best results for commercialism which is prone to manipulation?
- http://news.cnet.com/8301-17939_109-10242260-2.html (feedback on Wolfram)
- www.wisebread.com/the-student-who-created-a-pr-nightmare-via-wikipedia (censorship of Wikipedia)
- http://tech.slashdot.org/article.pl?sid=09/05/06/2237244 (phony wikipedia article referenced)