Add a "Does it work?" section to the README

[alxmrs]

Let's see how effectively we can calculate #36 (in time and compute costs) and add a summary of our findings to the README.

[alxmrs]

I'm open to other ideas for proof that this package does, indeed, work.

[batman00723]

Hey!
I'm a beginner and I want to contribute to this repo. Can you explain to me what exactly I need to do in this repo, and I am willing to contribute

[alxmrs]

Hey Aman. Certainly, I'd love your help. Check out #33. This PR is really stale, but the idea behind the "mara_join" script is still as relevant as ever. How about to close this issue, you replicate what that PR attempted to do in a new branch? Here are a few additional pointers for how to go about it:

• instead of ARCO-ERA5, try using EarthMover's copy of ERA5 in ArrayLake. This is a better fit for SQL-like queries. See https://earthmover.io/blog/icechunk-1-0-production-grade-cloud-native-array-storage-is-here#:~:text=Get%20Started%20with%20our%20Earthmover%20ERA5%20Sample%20Dataset.
• Structure the Python script as a "uv script" in an examples/ folder, see https://docs.astral.sh/uv/guides/scripts/. And, let's try to make the script look more like a SQL/DataFusion query rather what the demo branch is currently doing.
• If possible, I'd love to see use of a SQL aggregation in the script. Maybe, read more about the domains of the two datasets to see what could be interesting to calculate.
• Once the operation is working, let's try to measure its performance. You can try captioning time with this (https://docs.python.org/3/library/time.html#time.perf_counter) and/or profile the script via perf_tests/.

Let me know if you have any questions. I look forward to your contribution :)