You'll need to change a number of files:
You don't have to use this file, but ansible will require an inventory
file somewhere. There are two sections, [cassandra] for all the
machines that will have Cassandra installed, and [opscenter] for the
machine that will have opscenter installed.
This contains the cluster name, seeds, and opscenter 'stomp address' setting. Additionally, the multi-DC version has public IP addresses for each of the nodes; this particular playbook setup doesn't make any EC2 API calls so the public addresses have to be hardcoded into this file.
You'll need to rename these to dse_credentials.yml and email_credentials.yml and supply actual values for them.
You may have noticed that the cassandra.yml file in group_vars, above, doesn't actually have a lot of configuration in it. The rest of the configuration exists in this file. It's the default cassandra.yaml file that ships with DSE 4.6.6 with a few changes for the small amounts of templating currently done and a snitch appropriate to single-region or multiregion setups. If you want to override the defaults, you'll need to change this file.
In the future, this file is likely to become much more templated so you only have to make changes in one file.
Future work may include changes for e.g. RedHat, Centos, etc.
Datastax recommends using Oracle Java for production installs, but openjdk is simple to install and works well for demonstration purposes.
Amazon's EC2, by default, comes with a very restrictive set of rules determining what can and cannot connect to your machines. Since these playbooks do not currently make EC2 API calls, they cannot set up security groups for you. If your nodes cannot connect to each other, this is likely the culprit.