pull down to refresh

This is a good example of a topic where you can tell whatever story you want depending on which metrics you choose.
In this case, it's truly evident that data quality impacts reality. This happens with other topics as well, which is why I always say that you need to be very careful when analyzing graphics. As you know, graphics depend on the input data. I like to know how the data was acquired before drawing any conclusions.
reply
37 sats \ 5 replies \ @gmd 4 Jul
It seems like insurance companies would have pretty good data on this and they seem to be reducing their risk significantly.
reply
Do you mean they are downgrading the risks or that they are reducing their exposure to the risks?
reply
42 sats \ 3 replies \ @gmd 4 Jul
It seems like they are withdrawing coverage from a lot of areas like Florida due to hurricanes and fire risk areas in Cali...
reply
Yeah, but even there, you have to be careful about how to interpret that.
Monetary damages from natural disasters have risen a lot, but most of that is because of how much more stuff has been built in disaster areas.
Deaths and injuries related to natural disasters have declined, but that's mostly because of how much more we've spent on safer structures and better emergency response.
reply
69 sats \ 1 reply \ @gmd 4 Jul
No data, just hearing whispers/frustration from owners online whose coverage has either doubled or been completely withdrawn.
It seems weird to me to go from insurable to non-insurable unless the see something in the data.
reply
I'm not sure about California, but I recall hearing that Florida recently had a major insurance policy reform that sent premiums through the roof.
reply
Beyond the actual quality of the data, you also have to think carefully about whether it's reporting the most relevant information to the question you're trying to answer.
reply
You're absolutely right. There may be hidden agendas behind data reporting.
Data quality is crucial, but it's only half the battle. The other half is ensuring the data is relevant to what you're trying to understand. In other words, relevant data might not be perfect data, but it needs to be suitable for the specific problem you're trying to solve.
reply