I don't think there's any way to estimate that. It's completely dependent on usage - uptime, mostly, but also CPU, bandwidth, and IOPS (overage costs are added to uptime costs).
The good news is that AWS recently started billing by the second, rather than by the hour - at a minimum of 60 seconds per instance. However, you still have to pay to boot up, shut down, or while idle. You also have AWS Billing Limits, Alerts, and Reports.
If you start with a t2.micro, and do most of your work there in CPU mode (no acceleration) until you absolutely need it, and are careful to shut it down, when not needed, you'll cut your costs way, way down.
If you convert your t2.micro instance to a p2.xlarge ($0.90 USD / hour), you can try to ask for a Spot Instance, to save money... Regardless, be sure to shut it down -- you're still billed, by the second! If you inadvertently left an instance running for a whole month, it could cost you roughly $700.00 USD / month -- IDLE!
To avoid costss, once again,
- Shut it down, unless you need it
- Use a t2.micro, where possible
- Use a p2.xlarge Spot Instance, when you have to
- Use AWS Billing Limits, Alerts, and Reports
I hope that sort of answers your question, maybe.