Frank Kane's Taming Big Data with Apache Spark and Python

Warning - Spark on EMR is not cheap

Now, I do want to warn you that if you have been through my course on the MapReduce and Hadoop, we also used Elastic MapReduce during that and it was dirt cheap. The reason it was dirt cheap is because we were using dirt cheap instance types, which is basically the kind of computer that we're using in our cluster. But with Spark, by default, we're using something that's a little bit more expensive, actually a lot more expensive. We're using m3.xlarge instances. So even though we'll see that Spark runs a lot faster, bear in mind we're using much more powerful computers than we did in the previous course. Personally, I racked up about thirty dollars in Amazon web service charges developing this course. I ...