Remove transactions that has duplicate source, target, amount & category + the time difference between the transaction is less than 1 minute.

remove-duplicate-transactions.js

// Converts the time string into Date Object
const timestamp = _time => Date.parse(_time) / 1000;

/** 
 * 1. group-by given keys 
 * 2. sort the groups in ascending order for Time key
 * 3. filters duplicate entries with time difference < 60 seconds
 *
 * @param _array    Array     [required]
 * @param _function Function  [required]
 * @param _seconds 	Number    [required]
 * @return Array
 */
const groupBy = (_array, _function, _seconds) => {
	return _array
		.map(_data => ({
			key: JSON.stringify(_function(_data)), 
			timestamp: timestamp(_data.time), 
			data: _data
		}))
		.sort((_a, _b) => (_a.timestamp - _b.timestamp))
		.reduce(([_accumulator, _previous], _current) => {
			if (
				!_previous || 
				_current.key != _previous.key || 
				_current.timestamp - _previous.timestamp > 60
			) _accumulator.push([]);

			_accumulator[_accumulator.length - 1].push(_current.data);

			return [_accumulator, _current];
	}, [[]])[0];
}

/**
 * Finds all transactions that have the same sourceAccount,
 * targetAccount, category, amount & the time difference
 * is less than 1 minute.
 */
function findDuplicateTransactions (transactions = []) {
	return groupBy(
		transactions, 
		(_txn) => [
			_txn.sourceAccount, 
			_txn.targetAccount,
			_txn.amount,
			_txn.category
		],
		60
	).filter(_group => _group.length > 1);
}

findDuplicateTransactions(_transactions);

What is the usage of [0] at line 32 ?

As the name suggests, it's the accumulator that holds all the transaction objects. You can return just the accumulator array from the return array. This way you can avoid using [0] from line 32.