Wolfram Alpha and cities
Posted on June 8, 2009
Filed Under General
As you’ve probably heard, there’s a new kid on the search block: Wolfram Alpha. Search is a bit of misnomer actually. Alpha bills itself as a “computational knowledge engine” and although it’s still early days, it does offer a powerful tool – and something notably different from Google. It’s definitely not a search engine.
Alpha’s main advantage is that it contains large amounts of curated data, codified in a robust manner that allows for “computation”. Reasoning might be a better word. If you enter a query like “GDP US vs. France“, it knows that you want to compare the GDP time series, whereas “GDP US / France” works out the ratio between the two values. In other words, Alpha is a demonstration of the potential of the semantic web. If all data is labeled and the relationships between types of data explained in an ontology, then computers can use this information to manipulate data and answer pretty complicated user queries: like “how many UK Premiership footballers come from home towns over 3000m altitude?” (why you’d want to do such a query is another, perhaps unanswerable, question).
At the moment, many of the Alpha datasets are incomplete and so you can only do limited queries. But since I work mainly on cities, I wanted to check what’s available on this topic.
The answer is not much, but then not much is all that’s readily available anyway. If you do a simple search on a major international city – say Lagos – you get the following:
I thought maybe if you tried a US city, the results would be better since much of Alpha’s data comes from American sources. But nope – you just get the basic location, population, weather, time and elevation.
But you can do neat comparisons: check this out. Flight info!
Sadly, for those interested in urban energy, you can’t do much else. Type in “London energy consumption” and it thinks you want to see how long it takes to fly from London to Energy, Illinois (7 hours 40 minutes, if you’re curious).
So overall, a neat tool but still a work in progress. I guess if I was to draw up a wishlist of changes, I’d like to see more curated city data sets (from the UN and IEA’s World Energy Outlook, for example) and an option to set metric units by default!. (If this bothers you too, go here and vote for the feature. In the intro video, Stephen Wolfram mentions that they use IP lookups to contextualise the results – i.e. do you mean London, Ontario or London, England – but I can’t tell if I’m seeing imperial because they think the UK runs on imperial units or if they just assume imperial for everyone.)
Comments
Leave a Reply