Earlier this year, one of our users Position How Domo and Jupyter notebooks were used to perform sentiment detection, a form of sentiment analysis, in app reviews on the Android App Store.
I wanted to share some code with you on how this works. This analysis collects reviews and attributes from your favorite food delivery app, joins the information into a dataset, and creates a dashboard showing sentiment analysis for each review.
Kendall Louver shares her code I recreated the dashboard below in Domo. Jupyter notebook integrationIntegration with Jupyter allows you to develop more advanced analytics techniques and models and easily deploy them within Domo.
For example, instead of trying to capture sentiment out of the box such as a word cloud (which may differ from App Store ratings), to better understand user sentiment, use the natural language processing method. Run a review using the sentiment detection model found by Kendall. hugging faceyou can develop and deploy ML models with minimal effort.
You can also take advantage of scheduling options within Domo. For example, the dashboard below updates daily at 09:00 UTC. Other options are available and current Domo users will find the options familiar.
Additionally, it ingests data from the Google Play Store API via Python code written in Jupyter. This puts data from third-party APIs that don’t already have a Domo connector within the Domo environment.
Finally, we now also support sharing and collaboration in notebooks and integration with accounts within Domo.
The dashboard, code and steps below GitHub site.
