This is an area in which I’m very interested. I think it should be possible to build some tooling that would make it easier to construct such articles.
One dimension could be an abstraction layer over common datasets; if I could just pick and choose data sources that would make it easier to plug together. Currently data sourcing and cleaning is a bit opaque. This seems quite hard to solve in general, but if you could get a critical mass on a far interchange format, maybe public datasets would start using it. Otherwise someone in the community needs to curate wrapper layers on top of those public datasets.
Another aspect is the graph tooling/authoring; for this to really take off we’d need a “Wordpress for interactive articles”, basically some GUI tool or plugin to let non-technical users build these visualizations easily. I actually think this is pretty tractable.
The biggest hurdle i can see is more sociological; most people don’t think in terms of models. But at the margin perhaps better tooling could make “model curious” people better at this mode of thinking.
One dimension could be an abstraction layer over common datasets; if I could just pick and choose data sources that would make it easier to plug together. Currently data sourcing and cleaning is a bit opaque. This seems quite hard to solve in general, but if you could get a critical mass on a far interchange format, maybe public datasets would start using it. Otherwise someone in the community needs to curate wrapper layers on top of those public datasets.
Another aspect is the graph tooling/authoring; for this to really take off we’d need a “Wordpress for interactive articles”, basically some GUI tool or plugin to let non-technical users build these visualizations easily. I actually think this is pretty tractable.
The biggest hurdle i can see is more sociological; most people don’t think in terms of models. But at the margin perhaps better tooling could make “model curious” people better at this mode of thinking.