carefully designed
rich profiles
precise methods
costly
privacy-led aggregation
more details are less frequent
Decennial census (and census geographies)
Longitudinal surveys
Customly collected surveys, interviews, etc.
Economic indicators
Tied into the (geo-)data revolution, new sources are appearing.
Created for different purposes but available for analysis as a side effect
nature
resolution
quality
Potentially, much more detailed
in both space and time
There are different ways
Digital life (Twitter/X, Facebook, Wikipedia…)
Digital traces (record of digital actions (CDRs, metadata…))
Digitalised life (non intrinsically digital life in digital form (Government records, web…))
Bottom up (“Citizens as sensors”)
Intermediate (Digital businesses/businesses going digital)
Top down (Open Government Data)
Massive, passive
Nowcasting
Data on social systems
Natural and field experiments (“always-on” observatory of human behaviour)
Making big data small
Bias
Technical barriers
Methodological “mismatch”
Old or new?
Old and new!