Discussion
Something I always have trouble with when learning a topic is when the examples aren't any sort of concrete applications. Like showing how to use pointers with just some generic names, not giving any context to how they'd actually be used.
With this data analytics stuff, I do not like examples where they just generate a dataframe full of random numbers and then do plots with it. I feel like to actually understand the concepts you need something tangible to tie it to.
So for example, I took Kansas City's open data of reported potholes, crunched it down to be a "month vs. count of new potholes reported" format, and used that for the polynomial regression.
All the websites I'm looking at just have some fake data generated as examples.
Like, why is there fake data around time of day vs. driver MPH, and the predicted outcomes are like... going 113 MPH at midnight???
I also avoid example code for things people may not be familiar with. For example, I always HATED having to learn about a sport in order to do my PROGRAMMING HOMEWORK.
Don't make me learn baseball to do programming homework. Grr!
@Moosader i dunno, in a way that's kinda "real world" though, first step in programming software to do something is always researching and defining exactly what the "something" is and breaking down the logic i'd need to interact with it, though maybe that crosses the line a bit between software engineering and programming and might be out of scope for a class on a programming language.
@raptor85 think of it as a difference between
void Foo( int& bar )
{
bar = 5;
}
(what the fuck does this even mean and why would we do this?)
vs.
void Quadratic( float& x1, float& x2, float a, float b, float c )
{
x1 = ( -b + sqrt( b*b - 4*a*c ) ) / 2*a;
x2 = ( -b - sqrt( b*b - 4*a*c ) ) / 2*a;
}
@Moosader The pothole example gives you another useful question too: are you predicting road damage, or how often people report it? A change in reporting could change the graph without the roads getting worse. That would make a good exercise after fitting the curve.