Machine Learning for Corruption - serenata-de-amor

[flippakitten]

So I am on a data science journey at the moment, don't know why I just find the whole big data, stream processing world extremely exciting, it's like where has this tech been all my life.

I can think of a million and one use cases, of which 5 are probably real use cases but that goes with anything really.

Anyway, I came across this:
https://serenatadeamor.org/
https://github.com/datasciencebr/serenata-de-amor

Basically it's machine learning where you feed data about public expenses and it will find anomalies where corruption might be present.

Now, does anyone here know of the best place to get that kind of data? I would love to see how this performs for us.

I mean we have plenty of data on where corruption exists thanks to the ANC and Mr Mashaba in Johannesburg.
Also, against those R300 lightbulbs in PE.

 

[creeper]

Auditor General + Request for information.

I'm also on the machine learning expedition at present. Looking currently at converting unstructured data to structured data and then connecting the structured data to existing structured data. The rules for ML are: 1) lots of data and 2) good quality data.

 

[flippakitten]

Auditor General + Request for information.

I'm also on the machine learning expedition at present. Looking currently at converting unstructured data to structured data and then connecting the structured data to existing structured data. The rules for ML are: 1) lots of data and 2) good quality data.

Thanks for the info, I was hoping it would be easier but I guess not.

But yeah, fully aware of the amount of data you need, just to get a basic model going.