### What changes were proposed in this pull request?
This PR proposes to fix:
- the Binder integration of pandas API on Spark, and merge them together with the existing PySpark one.
- update quickstart of pandas API on Spark, and make it working
The notebooks can be easily reviewed here:
https://mybinder.org/v2/gh/HyukjinKwon/spark/SPARK-35588-3?filepath=python%2Fdocs%2Fsource%2Fgetting_started%2Fquickstart_ps.ipynb
Original page in Koalas: https://koalas.readthedocs.io/en/latest/getting_started/10min.html
### Why are the changes needed?
- To show the working examples of quickstart to end users.
- To allow users to try out the examples without installation easily.
### Does this PR introduce _any_ user-facing change?
No to end users because the existing quickstart of pandas API on Spark is not released yet.
### How was this patch tested?
I manually tested it by uploading built Spark distribution to Binder. See 3bc15310a0Closes#33041 from HyukjinKwon/SPARK-35588-2.
Authored-by: Hyukjin Kwon <gurwls223@apache.org>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>