Slashdot Mirror
The Physics of a Good Store Location
Roland Piquepaille writes, "In 'Atomic Physics Predicts Successful Store Location,' LiveScience reports that a French physicist has applied methods used to study atomic interactions for another task: to 'help business owners find the best places to locate their stores.' Pablo Jensen has used his method for the city of Lyon and is now developing software with the local Chamber of Commerce to help future business owners. Read more for additional references and maps of the city of Lyon showing for example the best locations to open a bakery, according to atomic physics." Jensen says that more research is needed to know if this method would work in other cities.
That's two Roland junk science articles that have made Slashdot today. I guess since people have stopped clicking links for him, Roland has to up the submission rates to the editors.
ONCE AGAIN, he is linking to the same zdnet blog that he has the last 4 times. Are you editors all dumb? Can you not figure out it's a ad-trap? I guess since we can't filter him, we have to make posts like this to bitch. Did I mention this is more junk science to ad rape us with? K...
So yea offtopic me all you want Roland (with your many accounts here), but since (as usual) this is a submission for ad clicks, I just want to say thanks for wasting my eyeballs again. Let's hope people realize this is crap before they post and click for him. Not that his links have anything to do with real science...which makes this whole submission offtopic.
I know people (usually the business types) who can go to an empty store location, look around for 3 minutes and tell you whether the location is good or not.
Even on the same street, stores on one side may be "live" while stores on the other side may be "dead".
This is the kind of thing Economists have been playing with for years.
I can't remember who, but some guy got some press for an article he wrote discussing how a lot more progress would be made if we threw researchers from disparate fields together to work on a problem.
By bringing their differing talents/viewpoints/knowledge to a problem, you end up with new and 'better' solutions.
[Fuck Beta]
o0t!
I saw once this tag used for this submitter. Let's use it from now on! (/me tags pigpile)
#
#\ @ ? Colonize Mars
#
Next in the pipeline - physicists work with news editors to produce useless but well-targeted articles.
Here will be an old abusing of God's patience and the king's English.
For Firefox users, this Greasemonkey script will hide stories submitted by Roland Piquepaille:
http://userscripts.org/scripts/show/5738
This is great when you create a model based on a single town/city/location and apply it to that city/town/location, but that doesn't mean it is transferable.
For example, there are many small town or villages in wine regions in the midwest U.S. where there are multiple wineries/shops/bakeries that are right next door to each other and do quite well. According to this model, many of those bakeries should fail. In fact, it is the shops farther away from the "main street" that have the fewest customers. The same in historical/heritage communities. In the Amana Colonies in Iowa - same thing - multiple bakeries beside each other with similar merchandise but all full of customers.
What is missing in this model is how different cultures view the shopping experience. In the U.S., we seem to prefer going to a single area and having a large number of similar shops. If we want to buy a car, we prefer to hit an area where there are numerous dealers so we can find a good deal. We would rather drive past a local bakery to hit WalMart so we can save a few dollars. Since most american families own a car, how the distance and time affect things is different than in Europe. We don't shop locally by default, which is why WalMart can kill small businesses for miles around.
Now, I'm not saying that this model is bad, but the locality, culture an demographics needs to be taken into account, so this isn't a generic model that can be applied everywhere, but a technique that other localities can use to create their own model.
It would also be interesting to see how large changes such as extended construction on a major road, or the construction of a new mall/housing complex/office building would affect the model.
Reading code is like reading the dictionary - you have to read half of it before you can go back and understand it.
We wouldn't need doors on the stores because we could just phase through the walls.
God spoke to me.
There's an intriguing discussion of geographic distribution of commerce in Emergence: The Connected Lives of Ants, Brains, Cities, and Software by Stephen Johnson.
It's a great read, and it gave me all kinds of nonlinear insights into How The World Works that I hadn't really thought about before. It also definitely made me want to bust out Sim City again.
Read the EFF's Fair Use FAQ
Predicting the locations of bakeries is one thing, but what about call centers, NOCs, corn fields, and cnc shops? These are all pretty disparate examples where many variables preclude their location. For example, a call center needs access to telephone exchanges and a populace to draw against employee turnover. A NOC needs to be located where there are a lot of big pipes, preferably a tall building. A corn field needs to be in a place where there is enough fertile land that will support a large crop. A cnc shop needs to be in an industrial area with a lot of space and big doors. Not even to mention economic trends that affect this business, suggesting that even the smartest engineers and mathemeticians can correctly using physics explain all of these (mostly) subjective variables is purely absurd. A bakery or jewelery store, maybe even a grocery store or gas station.... sure, I'll buy that. But the application of this seems so limited that especially considering the articles vagaries and the person who wrote it, I call shenanegans!
This has to be a problem similar to that of deciding where in a store to put products to optimize sales, and that must be a pretty sophisticated science by now. Do you put the sesame oil next to the other oils or the asian foods? Laundry soups near the entrance or in the back? So the question becomes - is this approach to the problem better than others that are out there? Hopefully they have already tested it against other models and shown that it does well - before subjecting some unwitting small business owner to their fancy new software.
> help business owners find the best places to locate their stores
Um, we already know how to find the optimal location for a store. You look for where there's a McDonald's, and you locate the store right next to it. Couldn't be simpler. _How_ McD's always manages to find exactly the perfect spot, I'm not sure, but I've yet to see one suboptimally located, so plopping down next to them should be a pretty reliable way to find a really good spot.
Cut that out, or I will ship you to Norilsk in a box.
The article makes only a few points, all of them long held as valid. Summarized very briefly, sampling is used to correct for spacial anomalies, categories are revealed with Potts algorithm, and potential is revealed with field theory. What is interesting is that the writer of the article appears unaware of the large body of previous work similar to this. Also interesting is the Slashdot obsession with quick judgement over a thorough read. It seems that both the writer and the Slashdot audience reveal our culture of modernity to be limited not as much by the capacities of our Engineers and Scientists but rather by those of our Librarians and Professors.