I started experimenting with the real estate ML project. It's just the baby steps, but as you can see, the first prediction (in blue) is incorrect since it has a negative value. However, after performing cost calculations and adjusting to find the right values for w and b, you can see that the prediction is much better.
By the way, the prices are in millions, and this is real data from my hometown. Although it's just 3 sets of data, it's a great experiment!