One of my first projects with Python was to create a desktop cleaner. It auto sorted all the files by type so any documents would go into Documents, any images would go into Photos, etc. I don't recommend running it on ALL of your folders (I didn't make it recursive) because that might actually disorganize your files. I used it to just pick up any stray files on the desktop.
Another fun Python project is an auto-clicker. You can record your mouse movements and repeat them with Python, allowing for Python to take control of the mouse, open files or programs, click buttons, etc. It is a nice tool for repetitive tasks. To take this a step further you can actually supply Python with images of certain elements, like a program's desktop Shortcut. This would allow your Python script to find the program's icon on the desktop even if it has moved to a different spot. For example, I could code a Python script to find the Microsoft Word shortcut on my desktop. I could set up the file to automatically open Word when it finds the shortcut. Similarly, I could have my Python script create a new file and then type into the document, paste information, etc. The image-search style of automation is helpful because recording mouse movements and clicks fails the moment something takes a bit longer to load or if something is out of position. Image search fixes this problem by forcing Python to look for what it must click on its own. Basically Python takes a screenshot of the whole display and then looks for any part of the screen that matches the image you've supplied to it. When it matches, it performs an action on the element - in my example, it would double click the Microsoft Word shortcut, click the File option once, and then click the New file option - all based on small images of each option.