Dear Alan,

thank you this is really great. It is much better to have the data as a resource object. In his post on the Wolfram Blog Stephen Wolfram describes the steps that need to be done to get data into a truly computable form. He describes a hierarchy of 10 levels to complete data happiness. I think it would be great if we could come up with a post here on the community to work everyone through all of these steps for an interesting dataset.

Tonight I will post something else with a truly remarkable data set. I think that it might prove to be a good test object. In that post I will not be able to go through all the steps; I will rather perform a quick and dirty analysis, but perhaps I/someone can follow up on that.

I am also delighted to see that the Machine Learning environment is again functional for OSX users in Mathematica 11.1.1. I am sure that it can be useful to gain insight into these datasets.

Thank you for your comments and Wolfram Data Repository addition!

Cheers,

Marco

Dear Sander,

thanks for your kind words. I was quite happy with the speed up I got from my first naive implementation.

The amount of data is truly amazing. It is quite frightening that there are so many crimes being committed. I think that one would ideally use mySQL to deal with that amount of data. Also, I wonder whether one can use machine learning to get something else out of this. I am happy to see that MMA 11.1.1 is released so that I get to use my GPU on the Mac again!

Any ideas on how to deal with the "entire" data set and/or how to make the model more realistic, potentially by using curated data?

Cheers,

Marco