Advantage
Saving time and costs through efficient questionnaire evaluation
Data
Open answers from more than 40,000 survey participants
Methods
radient Boosting Machine
Challenge
The OBI Group is the market leader in the German home improvement and DIY sector and, with more than 650 stores, is one of the most well-known DIY retailers in Europe. To continuously improve its product range and customer service, OBI offers its customers the opportunity to share their feedback with the company via an online questionnaire. The open-ended questions in particular provide valuable and often unexpected insights. However, in order to evaluate the answers to these questions, they must first be manually categorized before analysis.
Goal
With more than 40,000 survey participants, manually assigning responses to appropriate, predefined categories such as product selection, location, or value for money is very time-consuming. For this reason, OBI sought a solution for automatically categorizing open-ended responses.
Data Science Implementation:
Three Steps to Your Successful Data Product
Solution
In a first step, eoda processed the responses submitted by OBI. To optimize the forecasting accuracy, similar terms were condensed using a word stem function, misspelled terms were assigned to a suitable term, and meaningless words were excluded. Furthermore, relevant terms that frequently occur in combination with one another were grouped together under a meta-term and thus implemented into the model.
Following the data processing, a model was created using the Gradient Boosting Machine data mining algorithm, which automatically assigns individual customer feedback to the appropriate response category. This model was applied to the open-ended responses, and the results were compared with OBI's manual assignment.
Result
The model developed by eoda achieves accurate response classification. With its high level of accuracy, the model provides valuable support for efficient questionnaire evaluation. Responses to open-ended questions no longer need to be manually categorized; instead, they can be automatically assigned to the appropriate response classes. This significantly reduces the effort and time required while maintaining high quality.
Get started now:
We look forward to exchanging ideas with you.
Your expert on Data-Science-Projects:
Lutz Mastmayer
projects@eoda.de
Tel. +49 561 87948-370
