It’s not easy to keep up with the hunger for geodata. There’s so much that must be mapped. Not only geography, but traffic, pollution, light, sound, crime... There are so many customers for the data and expectations around the richness and freshness of that data is escalating.
With all of these sensor networks (mobile phones, cars, city networks, drones) what are we sensing? Who decides what data is worth collecting and where? What forces drive the direction of our data collection?
Collecting data is easy and cheap. The hard part is processing it and making sense of it. It’s very easy to take pictures with a 360-degree camera array. It’s very difficult to stitch them together. It’s easy to collect a point cloud and a 3-D mesh. You can do it with an old Kinekt camera from eBay. It’s hard and expensive to turn that point cloud into something that can be read and used. Who decides what data is collected and how it is processed?
Humans tend to conflate automation and impartiality, automation and accuracy. We seem to forget that humans are the ones building the algorithms and AIs and that human interests lead us to build algorithms that can learn to read one type of data or another.
Data is just data and machines don’t know things. The political motivations and storytelling power of a map is not changed when the map is built by automated means and it’s important for us to remember that the same critical thinking you can apply to a traditional map has to apply also to the automated digital map.