kafka-connect-elasticsearch VS kafka-connect-twitter

Kafka Connect is a scalable and reliable tool for streaming data between Apache Kafka and other systems. It allows to quickly define connectors that move data into and out of Kafka. Luckily for us, there is an open-source connector that sends data from Kafka topics to Elasticsearch indices.

I'm reading O'Reilly's "Mastering Kafka Streams and ksqlDB" to start learning Kafka, it was suggested for me on an ad by Confluent. Unsurprisingly it uses Confluent's software throughout the book. One of the first projects is a simple app that does sentiment analysis on tweets. The book uses kafka-console-producer and a sample .json file for the tweets, but for my app I wanted to read actual tweets. To do that I've been reading about Kafka Connect and looking at this repository, but I'm having a hard time understating how to best deploy this for my local setup. So far I've been using docker-compose.yml files provided by the book, which in turn uses Confluent's docker images for kafka, zookeeper, etc. As for this Twitter Connect repository, it seems the recommended way of setting it up is to use Confluent's platform and its CLI tool to automagically install it, which is fine, but I wanted to learn how things work under the hood (to some extend) and if possible not rely so heavily upon Confluent's software. Is it a good idea to just stick with Confluent and the book, or should I be reading a different material for a first Kafka project and working with a different kind of setup? Perhaps I'm getting ahead of myself trying to use Kafka Connect at this point?