2000 to 2009: The Rise of Social Networks and Economic Crises - The Technical and Social History of Software Engineering - page 304

Zillow has data on about 100 million U.S. homes and Yahoo had data on about
5 million U.S. homes. The combination of the two probably includes a majority
of all U.S. homes. The data include both current values and historical past values.
There are also images on the Zillow site, including aerial and satellite images.
The company was criticized when it changed its appraisal method in 2011 when
it changed both current and historical valuations. Zillow claimed increased accur-
acy, but many users were not convinced. Zillow says that accuracy improved from
matching actual sale prices from about 12% to better than 9%. Why historical data
changes is not clear.
The site now has a number of products and services, including searching for
mortgages, applications that work on smartphones, and a tool called Zestimate that
can predict real estate sales prices. Zillow also produces local reports for about
130 metropolitan areas. Zillow also introduced a kind of Wiki service by allowing
users to ask questions online of a community of appraisers and other users.
Zillow demonstrates the fact that computers and software now permeate every
aspect of corporate and government activities. Property appraisals for tax purposes
have been in use for more than 1,000 years. Prior to the computer era with large
databases and the World Wide Web, appraisals could only be based on a small
sample of very local properties.
In today's world, essentially every piece of real estate in the United States has
data available online, so national, regional, state, municipal, and even neighbor-
hood real estate data can be examined.
In theory, appraisals should be more accurate today than in the past, but the
use of software containing proprietary and secret algorithms raises serious doubts
about accuracy. Also, software applications have a distressing tendency to contain
bugs and errors and no doubt this is true for appraisal software as it is for other
kinds of software.
This decade witnessed a large number of niche companies filling in the gaps
around more conventional database and corporate applications. Some of the new
niches include predictive analytics, real estate appraisals, and even personal rela-
tionships.
Social networks began to go beyond normal everyday human contacts in terms
of frequency. By the end of the decade, some young people had many more friends
on social networks than in real life. More time was being spent tweeting and tex-
ting than talking to people face to face. Texting was becoming so common that it