Wolfram Alpha research secrets

Zack Stern
12 June, 2009
View more articles fromthe author

I was born 5 hours and 17 minutes before a full moon began to rise over Minneapolis. The sky was nearly cloudless, and the temperature outside was 46 degrees Fahrenheit.

These facts are admittedly just a party trick from Wolfram Alpha, the Internet’s new “computational knowledge engine.” I just told it when and where I was born, and it came up with those details. But the engine’s future—and its real power—comes from comparing all kinds data on its servers.

Wolfram Alpha leaves the googling to the search engines. Instead of relying on spiders to crawl sites, employees enter and scrutinize its data, applying both manual and automated methods. Its value comes from making comparisons, overlaying stock charts, creating tables and graphs that report similarities and differences in socioeconomic data, and otherwise helping you draw conclusions from intersecting details.

Someday, when sufficient data is in place, you could compare a favorite baseball player’s home run record versus the time of the game. The framework exists for this kind of detailed search, and you can already make similar queries on other topics.

You have to think differently to begin tapping into Wolfram Alpha’s abilities. Here are some tips on how to get started with this new kind of comparison engine.

Start with small queries. Though it has been in development for five years, Wolfram Alpha often misunderstands queries. If you encounter problems with your search string, begin small. Try starting with a search for a place—say, “California.” Already, you can see how the tool dynamically incorporates the current date, as it reports that the state joined the union on September 9, 1850, which was (as of June 2009) 158 years ago.

Then pick another term that will produce overlapping or comparative results. Try “California income.” Simple enough. Each search result includes a pop-up window that identifies its source, in case you ever want to dig into the origins of Wolfram Alpha’s information.

Now try another overlapping term, such as “California New York income.” Wolfram Alpha generates a simple table for comparing income in the two states. Now, you may begin to see its potential.

The site is admittedly young and is versed in only certain topics. Thus, a search for “San Francisco income” comes up empty. If you cut a search back to its core and Wolfram Alpha still has nothing to offer, that entire topic might be missing from its current database. Visit more of the site’s examples to see whether a similar subject is available.

See if Wolfram Alpha understands. After you enter a query, Wolfram Alpha provides feedback about its comprehension. If you enter several similar terms, such as company names, the Input interpretation line should show them all. If you’re modifying a search, such as “1984 Apple IBM,” the interpretation shows that you’re getting the stock closing price over that full year for both companies. If you want a single day, try updating your search to say, for example, “January 24 1984 Apple IBM.”

Use location awareness. Wolfram Alpha already knows something about your search before you type a single letter. By using your IP address details, it estimates your location, thereby making contextual results possible.

Suppose that you’re searching from Austin, Minnesota, about that town; the engine already knows to report those results instead of the ones for Austin, Texas. However, pop-up menus let you correct the tool in case you do want the other city—or if you want results for the first name “Austin.” (If you meet someone named Austin, by the way, there’s a fairly good chance that he’s about 15 years old.)

Your location plays a role in other searches, too. For example, if you type “7-4-2009” within the United States, the engine will treat 7 as the month of July. If you were in Europe, Wolfram Alpha would treat the 4 as the month.

Stick to the facts. Steer toward topics that can be distilled into facts: city population, movie box-office draw, and Web site visitor statistics, for example. Wolfram Alpha computes in factual terms. But watch the right sidebar for links offsite. If you want more narrative, jump to Wikipedia and other sources.

Export results to Excel. Wolfram Alpha’s results are impressive for a Web browser, but they don’t easily convert into Excel spreadsheet data or into database content. The tool exports strictly as a Live Mathematica file or as a PDF; click those links in the lower-right of the results.

Even text results are rendered as images, so you can’t copy cell data and paste it into Excel either. You can temporarily turn the images back into text; but if you do, they will lose their cell formatting. But do it anyway—here’s how to move data into Excel.

Paste the text data into Microsoft Word. Notice the many | characters? Use Find and Replace to convert them into Tabs. Click the Replace button in the upper right and enter | in the find field. Enter ^t in the replace field, and click Replace All.

Copy the text again and paste it into Excel. It will automatically distribute itself into a series of cells. If you’re using another spreadsheet or database that lacks this feature, simply import the tabbed text file; nearly all database apps recognise those tab breaks.

Wolfram Alpha Professional will have a direct Excel export feature. A company spokesperson told me that the paid edition will be “released in the next few months,” with a price to be determined. Another extra feature of the Professional edition will be the option to cross-reference your private data with the engine results.

Enter mathematical queries. Since it’s built on Mathematica, another computational tool, Wolfram Alpha is already an excellent math aide. It understands many equations and terms from chemistry, engineering, physics, and other sciences. You can even try typing in point values, such as 1, 1, 2, 3, 5, 8, 13 … (with that space and ellipsis), and the tool will plot the points and offer its suggestion as to the source equation.

Compare medical results. Using its medical and socioeconomic data, Wolfram Alpha can provide useful medical baselines. Suppose you just had a cholesterol test, as I did, and have no idea what the number result means.

Start with a simple query, “cholesterol 153.” (Note that “153 cholesterol” didn’t work as we posted this article—another example of the engine’s limited initial understanding.) That query yields a distribution across the U.S. population. Add your gender and age for more-specific results.

Get nutritional totals. Wolfram Alpha’s database stores details about food composition. Search for “banana” to get full results about a typical banana’s dietary contents. But try a search for “2 bananas and .5 Snickers bar” for results that add all of those items. Note that the “and” operator is important here; without it, the engine thinks that you want a comparison, not a sum.

Just get started. Wolfram Alpha catalogs even more topics, including colour theory, astronomy, music, and conversions. Just start playing with searches in an area that interests you to gain further insight into the engine. If you have trouble, visit the video introduction or examples page.

Wolfram Alpha is still young, but it’s already useful in many topics—and its value will only expand along with its database of facts.

Leave a Comment

Please keep your comments friendly on the topic.

Contact us