Thread by @jnordvig

It used to be 'I am not an epidemiologist'; and now we can flip to 'I am not a criminologist'. In any case, there is huge interest in the topic of the rise in crime in 2020 in US cities, for basic reasons of safety, and well as for political reasons.

Below, some basic charts...

Here, we project full year crime counts in major categories in NYC using the growth observed in the first eight months of the year, to allow comparison with previous (full) years.

There are major spikes in key categories (auto larceny: 59.6%, burglary: 42.9%, murders: 34.6%, as has been widely reported. But some other categories are falling (rape: -24.6%, felony assault: -3.7%)

It is surprising, perhaps, given commentary, that overall crime counts have not spiked (although admittedly, in a year with lockdown, it may be hard to do a simple comparison like that)

When I posted some basic numbers last week, I was asked for longer history (still working on that). For background, the charts here focus on relatively recent history, which does not include the huge decline in crime in the 1990s, as discussed here:
theatlantic.com/politics/archi…

Looking at data since 2000 only will make the recent spikes look bigger than a longer perspective (as illustrated in the four figures referenced in a past tweet)

https://twitter.com/jnordvig/status/1299751841794854914

The key test will be the crime rates in the final months of the year for NYC, when activity (presumably) starts to normalize, more people return to offices, inhabitants return from 'summer getaways' etc.

The trends in the coming (somewhat more normal months), relative to the months of lockdown/protests/ riots, will be the key. They will tell a more indicate picture of what a new steady state may look like. That will be a good time to crunch the raw data for a balanced take. END

indicate = indicative...(sorry)