|Company A – the challenges
||Less accurate because the correct information is affected by the irrelevant word which in bigger size. In addition, the diversity issue caused by employees in different country, working hour and culture was not reflected in cloud.
||Irrelevant word could be solved by applying bi-gram or tri-gram tokenization in future.By checking on the raw data, diversity did not highlight in company review. More data require to represent comprehensive insight about the company in future.
|Company B – the good things
||The cafeteria is not considered as a good thing in Company B.
||By checking on the raw data, the a few comments on cafeteria is good which the word cloud represented the reviews accordingly. However, it is a subjective judgment. It would be convincing if more data and employee feedbacks collected for development and evaluation.
|Company B – the challenges
||Limited parking is not an issue in Company B.
||Future check on raw data indicated that limited parking is experienced by contract worker. Hence, the word cloud represented this issue correctly but not specifying to contract worker. It could be solved by applying bi-gram or tri-gram tokenization in future.