110 Turing’s Test
frameworks. Dörr discusses a number of ways in which automated journalism is currently being integrated into media operations, remarking
that such companies are still “evaluating whether products match the
standards and quality of their brand and how users and journalists react”.
49
The most extensive example of such an experiment is the one
being conducted by Associated Press and Automated Insights.
AP and Ai
The beginnings of automated journalism were modest. In 2007, Robbie
Allen, an engineer at the networking hardware and software company
Cisco, set up an online sports network called StatSheet. As he later remarked in a post for Automated Insights, Allen was a fan of basketball
and started the company to keep track of – and publish – information
on basketball teams.
50
As well as providing detailed statistics for NCAA
basketball, StatSheet began serving data reports for college football,
NASCAR racing, and the NBA basketball league, as well as providing
access to its visualisation tools so that sports enthusiasts could generate
such things as historical and real-time reports of such things as injury
times or scores. Since the 1990s, most sports had begun to accumulate
digital data on events – building on analytics that, in sports such as
football, dated back to the 1950s and 1960s; a decade later, such data
were becoming significant enough to provide big data analysis, as in the
work of Coleman, Bothner, Kim, and Smith, and Waguespack and Salomon.
51
An earlier, famous example of the power of data to revolutionise
a sport had been set out by Michael Lewis in his book Moneyball, which
detailed how the Oakland A team made use of information to transform
their position in the baseball league. The ability to identify top players
in American football games or track the statistical likelihood of superior
performance during Olympic events drew upon highly structured data
that could also be output to graphs, charts, and reports as StatSheet
did – all without any human intervention. Once templates were coded
and set up, software drew information from the huge tranches of data
being made available online and published results that were drawn upon
heavily by companies such as Yahoo! and ESPN. While sports provided
an accessible market for data-driven publishing, in 2011 StatSheets
changed its name to Automated Insights, securing $4 million in new
funding to help it to specialise in “realtime content automation” in areas
such as real estate, finance, and weather reporting.
52
As with its initial venture into sports automation, at this stage, Automated Insights (Ai) was concentrating on areas where there was a
demand for information to be provided in a way that could be easily
digested by people, but where there were not enough writers available
to produce content. From November 2010 to September 2011 when it
changed its name, StatSheets had generated over 100,000 articles with
frameworks. Dörr discusses a number of ways in which automated journalism is currently being integrated into media operations, remarking
that such companies are still “evaluating whether products match the
standards and quality of their brand and how users and journalists react”.
49
The most extensive example of such an experiment is the one
being conducted by Associated Press and Automated Insights.
AP and Ai
The beginnings of automated journalism were modest. In 2007, Robbie
Allen, an engineer at the networking hardware and software company
Cisco, set up an online sports network called StatSheet. As he later remarked in a post for Automated Insights, Allen was a fan of basketball
and started the company to keep track of – and publish – information
on basketball teams.
50
As well as providing detailed statistics for NCAA
basketball, StatSheet began serving data reports for college football,
NASCAR racing, and the NBA basketball league, as well as providing
access to its visualisation tools so that sports enthusiasts could generate
such things as historical and real-time reports of such things as injury
times or scores. Since the 1990s, most sports had begun to accumulate
digital data on events – building on analytics that, in sports such as
football, dated back to the 1950s and 1960s; a decade later, such data
were becoming significant enough to provide big data analysis, as in the
work of Coleman, Bothner, Kim, and Smith, and Waguespack and Salomon.
51
An earlier, famous example of the power of data to revolutionise
a sport had been set out by Michael Lewis in his book Moneyball, which
detailed how the Oakland A team made use of information to transform
their position in the baseball league. The ability to identify top players
in American football games or track the statistical likelihood of superior
performance during Olympic events drew upon highly structured data
that could also be output to graphs, charts, and reports as StatSheet
did – all without any human intervention. Once templates were coded
and set up, software drew information from the huge tranches of data
being made available online and published results that were drawn upon
heavily by companies such as Yahoo! and ESPN. While sports provided
an accessible market for data-driven publishing, in 2011 StatSheets
changed its name to Automated Insights, securing $4 million in new
funding to help it to specialise in “realtime content automation” in areas
such as real estate, finance, and weather reporting.
52
As with its initial venture into sports automation, at this stage, Automated Insights (Ai) was concentrating on areas where there was a
demand for information to be provided in a way that could be easily
digested by people, but where there were not enough writers available
to produce content. From November 2010 to September 2011 when it
changed its name, StatSheets had generated over 100,000 articles with