Developer's
Corner
 How
Addresses are Validated via the Internet
Melissa Data recently launched several new features
to the Data Quality Web Service, including geocoding
for Canadian addresses, and multiple record batch
processing. Bud Walker, Melissa Data’s new product
manager, gives us the skinny on how the new features
work and how it all applies to the programmer.
Q: What are the benefits of Web services?
A: Web services provide a simplified mechanism to
connect applications regardless of platform or
language. They are based on industry standard
protocols with almost universal vendor support and
can leverage the Internet for low cost communication
and a variety of functionality. Web services are
self describing and can be automatically discovered
and I believe that our Data Quality Web Service is a
perfect use of this type of emerging technology.
Q: Is it easier to call the Web service rather than
a local object, since you don’t have to install
software?
A: A lot of programming languages actually have a
built in SOAP tool kit that encapsulates the Web
service automatically and creates the SOAP object
for you. Also there is no maintenance so once you
create the application to utilize the service there
is no updates like you have with a local object. So
yes it is definitely easier.
Q: What is SOAP?
A: SOAP is an acronym for Simple Object Access
Protocol, which is a way of communicating between
client and server using pure text. There’s nothing
binary. Think of it as a way of defining a set of
exchange rules for a Web service. SOAP makes it easy
to understand the different input and output
properties that are available to the service. For
instance, you can define what output elements you
require like address, city, state, ZIP, and latitude
and longitude. You can just look at the description
and know what kind of data we’re going to offer back
to you, automatically, without having to actually
pore through the manual. In that sense, it’s easier,
but it still initially requires a programmer to
implement the service natively in your applications.
Q: What is the speed of the Data Quality Web
Service? How fast do you validate records?
A: We hover between one and two seconds per record,
if you’re only doing a single record, which is one
of the functions of our Web service. If you’re doing
100 a time, via a multiple record lookup, we can do
that in three or four seconds – mostly because it’s
just sending one large transmission to us, which we
then process on our end and promptly return. It’s
not a whole series of calls and responses. That’s
why it’s a lot faster to do multiple records.
Q: Regarding the new announced features, what is a
multiple-record lookup?
A: As we discussed, we have two functions with our
Data Quality Web Service. One is a single record
lookup and the other is multiple-record lookups.
Single record is best used for Web entry forms or
call centers where addresses are being entered and
validated one at a time.
Multiple-record lookups is an efficient way to clean
up your existing databases. For example, if you’re
storing them all day, you can send them in batches
at night, clean them up and then store them in a
data repository.
Q: Why would you want to use multiple-record batch
processing?
A: Speed for one. In multiple records, you’re
sending 100 at a time so it’s more efficient and
faster. If you have one lookup, it goes through the
Internet, then it comes to us and we send it back
verified. For 100 the connection is opened one time,
and we process this quickly and send it back again.
Usually it is around one second more for 100 records
as opposed to one record.
Q: I always thought that the more records you have,
the longer it would take than if you just had one
record to process.
A: No, because of the travel times. Once we open up
the SOAP envelope, and take out the addresses, we
just let run them in batch onsite and put them back
in the envelope clean. But that process has to be
repeated every time for a single address.
Q: How do you determine gender with your name
parsing?
A: We have a large first-name database that codes
the most common male names and the most common
female names. In the case of neutral names like
Chris or Pat, we will just return the name as
Neutral and we will assign an Unknown status if we
can’t determine the gender based on the first name.
The database has 13,705 names.
Q: What’s new with name parsing?
A: We are in the process of rolling out with a brand
new and improved Name Object by the end of August.
The feature list is extensive. You are able to have
customizable name databases, and customizable vulgar
word databases. We are also adding an advanced
parsing engine so multiple names like Mr. Darren and
Mrs. Tabitha Stevens will parse and standardize.
Really it’s a whole new step in name parsing and
gender coding. This could really help your business
get a handle on your customer base and help you
pinpoint your marketing efforts.
Q: Can you describe the new Canadian latitude and
longitude feature for the Data Quality Web Service?
What does it do?
A: Prior to this, we’ve only be able to verify
Canadian addresses. There was no latitude or
longitude like we had for the U.S. So, if you wanted
to have that geo-coding information, you’d have to
look elsewhere. In Canada, the postal codes are very
precise, and the latitude and longitude based on
that postal code is also of a very high resolution.
So, now after we validate any Canadian address, we
return the latitude and longitude.
Q: What are the practical applications for the
geocoding programmer?
A: Well, the same with any geo-mapping, geo-spatial
mapping – you can figure out where your customer
base is and you can present them visually on a map
and determine where they are. There are all kinds of
bonuses to just knowing where your customers are.
For many companies like shippers and marketers this
is an invaluable tool.
CNN, for example, used the U.S. latitude and
longitude of our GeoCoder Object – when they were
covering the last election. That way a person
visiting their website was able to plug in an
address and determine who their senator and local
politicians were and nearest voting place. But with
the addition of the Canadian latitude and longitude,
this type of lookup is now possible for Canada as
well.
Q: Well, thank you so much for your time Bud.
A: You’re welcome.
|
|
|
