How to conduct my first data-analysis study? I'm trying to improve my understanding of a few basic concepts in statistics (like statistical significance, the p-value, etc.) and to achieve that I decided to do a study myself (trying to learn through doing).
I collected a data set about my chess games that includes my level of alertness, the number of blunders I made in each, the result, etc. My hypothesis is that the more tired I am, the more blunders I make.
I'm kind of clueless of where I start and that's why I'm asking for your help. Is there a good primer book (or a series of articles) that helps me devise the study? Is there a platform where I can plug in my numbers that helps me arrive at the relevant values?
Thank you very much in advance,
Balint
 A: BruceET gives a lengthy answer, but I believe a better model can developed than a simple test of proportions.
Let $b_i$ be the number of blunders in game $i$.  Let $n_i$ be the number of moves you make in game $i$.  As an avid chess player, I know that blunders are just a fact of life, and when games are lengthy you are more likely to make more simply because you have lots of opportunities to blunder.  Hence, we need to account for how many moves you make in each game.
This can be done with simple Poisson regression.  Our model will be
$$ \log(\lambda_i) = \beta_0 + \beta_1x_i + \log(n_i)$$
$$ b_i \sim \mbox{Poisson}(\lambda_i) $$
Here, $\exp(\beta_0)$ is going to be the expected number of blunders you make in a game where you make $n_i$ moves. The term $\log(n_i)$ is known as an offset term and does not have an associated coefficient. As you make more moves, the expected number of blunders will naturally increase.  The variable $x$ is for your measure of alertness, and so $\exp(\beta_1)$ will be the factor by which the number of blunders changes per one unit change in alertness.
