Interesting stuff. we have something similar using SQL agent jobs, and SSRS.
Unless I missed it, the prediciton doesn't seem to have methods for predicting impact of reindexing and check database activities. Do you have metrics for establishing the peak DB size, or else of predicting the space required for that? We find that it is during the maintenance that our space alerts (85%, 90% etc.) most commonly trigger.
I've been using the process described in the article for the past year and haven't had the need to go any further than the data file. You could of course create a similar process that looks at table and index usage, but this would be slightly more complicated. The situation you describe with free space is somewhat trickly for instance I can point to several databases that basically are one giant table. These databases will need more than 15% or 10% free space to reindex. There really isn't a good rule of thumb for percent free space in database. However, by looking at the growth of data files over time which includes used space, as long as you don't shrink files afterward you should be fine.
Forecasting storage capacity is bit like horseshoes, you want to get things close enough, but anything more complicated may not be worth the effort. I've found I'm able to catch databases before they autogrow and able to detect abnormal growth conditions. But, I also find some false positives such as when a new database is first deployed and the data points are small. I'll go through the list of databases with less than 90 days left and research each one individually. This list is typically small, less than a couple of dozen in an environment with 1000+ databases. The ones that need to increase in size I will do so manually and the ones that are false positives I will ignore after research.