To handle a stream of incoming data, these numbered steps (as documented in the Flink documentation) have to be done programmatically:

1. Get the StreamExecutionEnvironment object. StreamExecutionEnvironment is the basis for all Flink data stream execution programs. The following code block shows one way to get this object created in Java:

 final StreamExecutionEnvironment env = StreamExecutionEnvironment . getExecutionEnvironment( );

2. Load the initial data from the appropriate source.
3. Do the necessary transformation on this loaded data. The full set of transformations available can be found in the Flink documentation at this link: https://goo.gl/YI82xY.

4. Specify the destination where the results after transformation has to be kept.
5. Trigger the program execution.

We wouldn't want to replicate the Flink documentation here; rather we would say that you go to this link to have a deep understanding of the DataStream API in the Flink documentation: https://goo.gl/NS69SK.