We used the publicly available data of US traffic violations, uploaded the whole data set to Aito and analyzed if there are interesting facts to find in the who/why/when traffic violations occur.
For this example, we chose to use the Predict API.
The visualization below predicts the timing when certain violations are most likely to occur – based on the car brand.
Indeed, the results don’t show too much differences with the different car brands or violation types. And while we hoped to find very surprising facts (e.g. BMW drivers get much more violations during the night then KIA drivers), the results show that the US legal system is actually working as it should – handing out traffic fines in a quite uniform way, regardless of time of day or car brand.
We, on the other hand, succeeded in our example: finding correlations within a big data set without having to invest a huge amount of time and money in data preparation, data model building or training data.
But maybe you can find some interesting correlations with other API endpoints?