Satellite data aims to improve location searches
Satellite data, road and transport timetables are helping to make travel-time searches more accurate.
Most search engines plot travel times "as the crow flies", which can produce unrealistic estimates of how far it will take to reach a location.
Two British projects are changing that by basing estimates on timetables, road data and more accurate maps.
But the efforts of both have been hampered by bad data sets that made it tricky to combine information.Local hero
"About 40% of all web searches are for geographic information," said Charlie Davies, founder of mapping firm iGeolise. That percentage rose to more than 50% when people searched from a mobile, he said.
Before now, he said, most results were derived by the straight-line distance between two points.
IGeolise, and others such as Mapumental, were starting to produce results that were more humanly relevant, said Mr Davies. These no longer ignored important human factors such as the position of bus stops and stations, the frequency of trains, buses and trams and key details of UK roads such as the distance between junctions.
Combining these data sources with average walking speeds and driving speeds produced much more accurate results, he said.
It would mean that instead of just searching for locations within a mile of a person's house, they could be much more specific with their search parameters. They could, he said, seek locations that were 45 minutes away by car or could be reached in a 45-minute commute given that a person set off at 07:00 every morning.
In early November, iGeolise won a European Satellite Navigation Competition for its mapping work.
However, said Mr Davies, getting at the data that could help produce these types of results had not been easy. While many public authorities and agencies make their data available it was far from easy to combine it, he said.
"There's a lot of data and it's all in different formats," he said. Data from different sources even when it was about the same kind of thing, such as bus times, was rarely preserved in a file in the same way. Before it could be used, he said, it had to be cleaned up and standardised.
Tom Steinberg, head of the MySociety project that is behind Mapumental, said it was often public transport data sets that was not well preserved or presented.
"The data that is put out has often not been collected without much or any thought that it would be used by third parties," he said. "Typically this leads to the data containing ambiguities or apparently internal contradictions that take a lot of time to clean up." Mapumental would soon be letting the public play with its mapping system, said Mr Steinberg.
"We're very grateful to have the data," he said, "but it is clearly early days."