# Nonsense Data with Wolfram Alpha + Mathematica?

I am trying to learn how to interface with WolframAlpha. When I try using the Mathematica interface I get nonsensical data. I don't know if this is a bug or if I'm doing it wrong.

My goal is to extract the elevation around the Austrian Province of Upper Austria. Perhaps to assert that Upper Austria is not actually the highest Austrian province, if this doesn't work I will attempt to challenge the status of Lower Austria.

My Wolfram Alpha reference query is:

http://www.wolframalpha.com/input/?i=average+elevation+around+upper+austria+Austria


My Mathematica code is:

WolframAlpha["average elevation around Upper Austria, Austria \
(state)", {{"ElevationStatistics", 1}, "ComputableData"}]


which gives

{{"average", -(27896275/24384)}, {"lowest", -(3662500/
381)}, {"highest", 2585000/381}}


which can't be right by any measure. What's going on?

-
It works better with plain Austria. The Upper seems to mess things up. You might want to specify latitudes instead... – Yves Klett Oct 10 '12 at 5:55
Given the proximity of this region to the ocean, I wonder if it's simply taking ocean depth into account? – Mark McClure Oct 10 '12 at 13:19
@MarkMcClure the elevation profile looks about right for that (if you mean the Yemen, Austria is another thing entirely). – Yves Klett Oct 10 '12 at 13:23
@YvesKlett Yes, exactly. I tried a few more coastal towns ("elevation around miami, fl", for example) and saw very similar behavior. – Mark McClure Oct 10 '12 at 13:29
And obiously, "elevation around 48°E 14°N" does seem very familiar... – Yves Klett Oct 10 '12 at 13:35

The bug is probably with Wolfram Alpha, not Mathematica, since the Mathematica query and the Wolfram Alpha website results are consistent.

(Edit: Bug is fixed as of 2012-11-26, I got a kind notification from the W|A team in response to my earlier feedback)

It looks like they swapped longitude with latitude for Upper Austria (it works with good ole Austria though). What you presumably get is a piece of the Gulf of Aden instead.

Image from the query:

48°E 14°N:

Anyway, I am not sure this really pertains to this site if it is a Alpha Bug?

Edit by belisarius

Aligning and animating Yves' ansatz:

-
I think its relevent as it is correct when acessed with the Woflram Alpha Interface. – Mikhail Oct 10 '12 at 6:52
@belisarius epic alignment! +1 if you post that as answer. I wonder where the large angle difference comes from. – Yves Klett Oct 10 '12 at 8:54
@MarkMcClure: Would you say that if at some time a file system bug "leaks" into the Get command, then general file system questions should get on-topic here as well? – celtschk Oct 10 '12 at 15:23
@celtschk Certainly, the process of figuring out where a bug that leaks into Mathematica should be fair game here. If the OS happened to be written in Mathematica by the makers of Mathematica, then all the more so. – Mark McClure Oct 10 '12 at 15:30
@YvesKlett The angles come from me not looking at the Google Earth bearing. If you align the North, you get !Mathematica graphics – belisarius has settled Oct 11 '12 at 16:52