Dataset of the Day: Starbucks Closure Data
July 18th, 2008by Bill Greer
Sometimes it seems like there is a Starbucks on every corner, and sometimes it’s true! It looks as if they have finally reached their saturation point and are now closing 616 stores throughout the United States. This Finder! dataset shows the locations of the closures. We also uploaded a dataset that shows the almost 9,000 Starbucks locations around the Globe. With this point data, you can see that many of the locations being closed are very near to other Starbucks locations. Perhaps it makes sense to close stores that would cannibalize your own market however, there are many other ways of looking at the problem. We aggregated the data out to the Zip code and to urban areas. In case you were wondering, here is a sneak peak of the locations most impacted by the closures:
By Zipcode
1. 89108 Las Vegas, NV (5)
2. 63103 St. Louis, MO (4)
3. 77102 Houston, TX (4)
4. 92101 San Diego, CA (3)
5. 63102 St. Louis, MO (3)
By Urban Area
1. Dallas Fort-Worth Arlington (25)
2. Los Angeles-Long Beach-Santa Ana (22)
3. New York-Newark (22)
4. Chicago (18)
5. Las Vegas (15)
Lastly, we decided to map out some of the Starbucks locations with a competitor to see if perhaps that played a role in the closure decisions. Below is a map from New York to Philadelphia showing Starbucks locations (transparent green dots) and Dunkin Donuts locations (transparent magenta dots). The black dots are Starbucks locations which are on the Closure list.
Popularity: 20% [?]
July 19th, 2008 at 12:04 pm
I think you should post more often, I have enjoyed this so far. Added to my reader. SusanO
July 19th, 2008 at 3:50 pm
Sorry, but I do not agree totally with you. I am afraid that ALL affected are impacted. MOST is foe everyone …
July 23rd, 2008 at 11:48 am
I was the original harvester of the 9000 Starbucks locations. I’m pleased to see the dataset being used– very interesting.
August 10th, 2008 at 7:33 pm
Starbucks just closed 61 of its 84 stores in Australia.
List of closures:
http://www.starbucks.com/australia/closures.pdf
Remaining stores:
http://www.starbucks.com/australia/stores.pdf
August 11th, 2008 at 10:27 am
Thanks Tim - we’ll get it geocoded and up into Finder! shortly. It will be itneresting to see if the patterns are the same as in the US.
best,
sean
August 11th, 2008 at 2:56 pm
The Starbucks closings in Australia is now up on Finder!
http://finder.geocommons.com/overlays/3875?page=
Thanks Tim!
Emily
August 12th, 2008 at 10:23 am
Is there an easy way to do that? I am a noob, but willin gto save you guys the hassle next time.
August 14th, 2008 at 8:16 am
Hi Tim,
You can do visual comparisons just looking at the data for both countries at the same time, and seeing if it looks like the closures are in the big cities or else where. Alternatively you can do spatial correlations to see if there is a mathematical relationship between the closures and other variables. For instance in the usa there is a dataset in Finder with the total number of Starbucks in metro areas so you could see if they were just closing stores based on saturated markets. You could go more exotic and bring in home foreclosures or gas prices and see if there is a relationship there. We’ve not exposed this type of functionality in our system, but all the data is there to do it off line. You could use a GIS package for the spatial dimension or just excel for non-spatial regression or pearson correlations.
best,
sean