Unsupervised Learning on Customer Reviews Dataset

A simple project to perform Association Rule Mining and K-means Clustering.

1. Excel Reader Node

Importing the hotel dataset to the workflow.

2. Duplicate Row FIlter Node

Filtering duplicate rows in the dataset, to remove repeat data as the total number of records amounted to 52,417. But, it consisted of 9 duplicated datasets containing 5824 records in each duplicate.

Therefore, applying this node enables the removal of the duplicates to ensure that the dataset is cleansed. As a result, there were 1012 records.

3. Excel Writer Node

Exporting the cleansed dataset

Excel Reader:

Select Attributes:

FP-Growth:

Create Association Rules:

Based on the results, we can observe that most of the terms associated together in the association rules are such as “satisfied”, “room”, “pool”, “large”, “good”, “stay”, “staff” and “excellent”. Hence, we can derive that customers are mostly satisfied with good staff service.Otherwise, we can also infer from the association rules that customers enjoy hotel facilities such as the large pool in association rule #4 as an example.

Hence, I would recommend that customers are usually looking for an excellent and satisfactory experience in hotels with large pools, good staff service, large hotel rooms.

1. Import the cleansed hotel dataset to SAS Viya

2. Select clustering as the object

From the results shown in the statistics table, we can see that all of the customers are under the country region of SouthEast Asia. Cluster ID 1 of customer travel type, couples, has the highest star rating among all the clusters with an average rating of 8.5 stars. Slightly below is cluster ID 4 which are Families with Young Children with an average rating of 8.44 stars, right after is cluster ID 5 which are Solo travellers with an average rating of 8.32 stars. We can understand from this evaluation that Couples, Families with young children, and Solo Travellers are almost on par.

Based on the model evaluation, it is recommended that the hotel set their target audience towards Couples, Families with young children and Solo travellers when marketing the hotel. They can create promotional bundles to cater to group customers like couples or Families with young children. For example, children in families are allowed to have free admission to the hotel.

Related posts:

Datetime in python functions to process the datetime type in the program we want to make, for example we want to know the time when the program is run or even the day before and after the datetime…