Parse JSON with Async Iterator

package.json

{
  "name": "async-iterator-with-json-stream",
  "version": "0.1.0",
  "description": "Use an async iterator or generator with streaming JSON.parse",
  "main": "index.js",
  "scripts": {
    "test": "echo \"Error: no test specified\" && exit 1"
  },
  "author": "Matt Morgis",
  "license": "MIT",
  "dependencies": {
    "JSONStream": "^1.3.5",
    "request": "^2.88.0"
  }
}

parse-json.js

const request = require("request");
const JSONStream = require("JSONStream");
const stream = require("stream");

// Needed to bind JSONStream with async iterator
const passThrough = new stream.PassThrough({
  objectMode: true
});

const jsonStream = request
  .get(
    "https://s3.amazonaws.com/philadelphia-parking-violations-raw-data/parking_violations_2017.json"
  )
  .pipe(JSONStream.parse("rows.*"))
  .pipe(passThrough);

async function run() {
  let counter = 0;
  try {
    for await (const chunk of jsonStream) {
      counter++;
      console.log(chunk);
    }
    console.log(`CALLED ${counter} TIMES`);
    console.log(" ### Done ###");
    const used = process.memoryUsage().heapUsed / 1024 / 1024;
    console.log(`The script uses approximately ${used} MB`);
  } catch (e) {
    console.log(e.message);
    process.exit(-1);
  }
}

run();

Can you explain why this PassThrough was needed?
How come the JSONStream outputs parsed JSONs (i.e. actual objects) and we still need to convert them via an objectMode stream?

@johnib Explanation here: nodejs/readable-stream#333 (comment)