It's Not Just About The Customers

Learn Aster

For businesses, customer feedback is very important. It is all about how well you know your customers and how you can keep them happy. But what about the employees? Studies have shown that happy employees are, generally, more productive and motivated. By using Aster, we can visualize the association of the Net Promoter Score (NPS) with words in employee comments to give direction on areas that can be improved. Text input (employee comments) and structured data (1 - 10 ratings, Yes/No) would be combined to leverage Aster text functions. Datasets such as: public websites (like Glassdoor), surveys within the organization, or even social media can be used to better understand what keeps employees happy.

Understanding The Employees: Opinions, Concerns, and Suggestions

On this example, we will use text from the internal survey of Acme Corporation. We want to look at the words that employees use most frequently to describe Acme. These provide direction on which areas Acme need to improve on. We will utilize SQL-MR functions (cfilter and ngram) to perform the association and use Aster AppCenter to visualize the results.

To keep it simple, we will only use one question from the Acme internal survey as an example and create a dataset which includes employee_id, overall_nps, and the answers. The question is "Why are you likely or not likely to recommend Acme as a place to work?"

Data Transformation

Ngram (1,2 and 3 gram) is used so we can better visualize words and associate them with the overall NPS. It is also necessary to remove stop words in order to reduce the noise on our visualization. Before using cfilter, we need to create a column that contains both the ngram and overall NPS then use it as input for cfilter.

Analysis and Visualization

We are now ready to execute cfilter to get the ranked list of words associated with promoters and detractors. Just filter out the col1_item1 from the output that only contains "Promoter" or "Detractor" as values.



Typically, sigma graph is used when visualizing cfilter output. In our case, it involves a lot of dragging around nodes. Additionally, we want to know what words follow adjectives to give more insight on what employees say.

With data manipulation and a little creativity, you can visualize the results in a much more elegant tree diagram using AppCenter:


By using text analytics on employee survey data, we can understand which areas within the organization can be improved to maintain a great working environment. Management can easily run the employee survey data on the app to generate insights so they can have a better understanding of their employees.