USE THIS: http://turbowarp.org/394975460 Press "SPACE" to start island generation and reset island generator Use the sliders to modify the island shape. The blue box shows the boundaries of the distribution and the red box shows where the island will be drawn The standard deviation variables will change the width/size of the island and the mean variables will change the position of the island. The noise variable will change the amount of noise used in the island generation. It will make the island look very rough, so I recommend making noise either very low or 0. It does not affect the distribution The detail variable will change the number of points generated and the scale variable will change the scaling of the data/island visualization. NOTE: Increasing scale and/or detail will greatly affect performance. Even if you have a decent computer, I would recommend not increasing both of these variables at the same time. If you want to increase one, decrease the other. For example, if you want to increase the scale, decrease the detail. Otherwise, it's gonna take forever to run. If you want to see a heat map of the data, hold q on the menu and set @island to 0. This was just a test feature, but the output is cool to look at
Explanation on how this works here https://scratch.mit.edu/projects/395119010/ Simple island generation using the bivariate normal distribution. Probably not the best way to do it, but whatever - it's cool regardless. To generate the distribution, I used Monte Carlo sampling under a standard normal pdf. It then scales these values with the inputted standard deviations and means. Next, it checks each pixel and takes certain distance parameters, such as average distance and min distance, to calculate the pixel colors. Pretty cool stuff Idea for project and all code by yours truly :P