The Core
Why We Are Here => Web Development => Topic started by: dogboy on February 10, 2011, 01:03:48 PM
-
I need to find a list of chain restaurants, coffe shops, and bars... And Starbucks counts as 1,not 10000. Then I need to go down the list until i find places like 'joes tavern' and each one of them may be different, but most likely they will count as one too.
The goal is for you to tell me the name of a place and 90% of the time, it will be on my list, and I'll be able to tell you whether it is a bar, coffee shop, or restaurant. And note I said 90% of the time, not that I need 90% of every possible place in existence. I also just really need the name and it's type... I don't need contact info, I just need to identify them.
I just started searching but I thought there had to be someone in here that might be able to point me toward some good sites for this type of thing. I found a couple high in search, but I'm wondering if I'm missing the obvious like somekind of industry association or public data.
I found these after about 10 mins. Know of any sites similar or bigger?
http://www.odditysoftware.com/page-datasales9.htm
http://www.databasepool.com/index.php?l=page_view&p=database_all
-
I came across this a few years ago when I was looking into something similar -
http://www.fpsgame.com/us-databases.html
Can't vouch for them as I never got around to the project but they might be an option for you...
-
>they might be an option
...they might be a good one:)
http://www.fpsgame.com/us-bars-and-taverns-database-1.html
...But I just re-thought all this and I'm wondering about going more 'live'/2.0, and am looking at emerging technology like:
http://simplegeo.com/products/places/
...an example of their tech is here:
http://m.whatser.com/#query=Bar&view=search
...and despite how easy it would be to purchase awesome (but dated) info, I'm seriously wondering if this approach isn't a much better solution for my (mobile oriented) problem.
In the meantime I'm looking at both, so if anyone has any other suggestions, I'm happy to cherry pick:)
-
ok, back to trying to figure out how to amass a database.... eventually we'll go API, but not in this version.
so another shout out: anyone know of who might provide a complete listing of restaurants and bars in US?
-
Yellow pages have a mobile site with nice clean HTML/interface ... you can search by ZIP code & distance
http://m.yp.com/94102/search/restaurants?geo_locations=94102___San+Francisco%2C+CA
You probably already knew that, but just in case...
-
I have people peeking under the hood... I need legal use.
Funny though I just posted the same thing on FB... a YP database (or API is what I need:) I'll look harder but it's owned by ATT...
-
I'm not sure the Yellow pages will include all the ma and pa restaurants.
Can you get a APIs from places like Yelp, City Search, ChefMoz and some others and aggregate the results like a metasearch?
-
If you need legal then paid may be the way forward.
A good place to start would be Gib Olander at Localeze.
Not sure what the cost would be for verticals as we do full IYP (and source our data from elsewhere due to advertising tie-ups). He seems like a good guy so if he does not have what you need, he may be the man to ask where to look next.
-
We spoke to a lawyer and they said if the information is something you can find in a phone book (business name, street address, phone number, etc.) then it is public domain.
So if you collected the data from a site, it would be legal to use that data provided you didn't break their TOS by collecting it.
-
Localeze is 50k-100k/yr with monthly updates, depending on what I want... my pockets arent that deep.
I pointed out this link the first time around but a buddy showed me there were a few more here I missed:
http://www.odditysoftware.com/download_databases/12_food-and-drink-related-data_2.html
>if the information is something you can find in a phone book
the trick is how do you get it on a big scale without violating tos by scraping... scan the paper book?:)
>Can you get a APIs from places like Yelp, City Search, ChefMoz and some others and aggregate the results like a metasearch?
some, but yelp also uses localeze and city search hasnt called me back,,,
I guess I found the low hanging fruit. I wasn't sure if you guys might come up with govt liquor license data or something like that. I think I better to stick to paid dbs for now, for cost's sake then expand from there...
thx, guys
-
With you stating that you need to only say whether a business is a given type it may be possible to query a search sevice such as Yahoo BOSS as you get a query - it depends on the usage you envisage.
Here's a possibility...
Search for this
"Business Name in Double Quotes" AND ("coffee shop" OR "bar" OR "Restaurant") "placename if required too"
You should then be able to use the snippets (and the knowledge gained from the URL's returned) to make a determination.
Even better, you could use the "sites" parameter in BOSS to search for a business name in say 30 vertical sites (10 for each vertical, each one very specific to only one of the business types), this would then give you the ability to see which verticals there are matches on (and, by using the returned results count, see if the search is likely to be specific enough or not).
I'm not sure what the legality of this is, so you'll have to check.
There is the issue that it's not going to deal with poor spelling, but you could probably get around that with a bit of thought.
-
Of course, if legal, you could keep a local copy of the outcomes you reach and then you'll qiuckly build a database that has commonly searched for business names in those areas - thus improving speed.
-
I'll copy and paste all this... much obliged:)