A project for Power BI hackathon.

Input data: dataset with posts and comments from the PowerBI Community Facebook group

Task: to do something useful with the dataset 😊

Requirements: must be a Power BI solution

Idea: to create knowledge base:

• detect if the post is a question
• mark questions as Solved if there is an answer to this question in comments.

Solution:

• created a dictionary of question words such as “someone”, “dealt with”, “please”, “need help” etc with values for “question probability” scoring
• created a dictionary of thank you words such as “thank you”, “thanks” with values to suppose if the question was answered
• developed Power Query function to detect these words in posts and comments
• created a dashboard:

Result:

1. Simple scoring model to detect questions
2. We’ve won the hackathon 😊

Links:

https://github.com/pavelveret/hfd4-krasivo — code examples + .pbix file