Domo + Jupyter for the Window Into App Sentiment

 / Domo + Jupyter for the Window Into App Sentiment

Earlier this yr, one among our customers made a publish about how she used Domo and Jupyter Notebooks to run emotion detection, a kind of sentiment evaluation, on app critiques from the Android App Retailer.

We determined that we wished to share with everyone some code on how this works. For this evaluation, we gather critiques and attributes from a few of our favourite meals supply apps, mix the knowledge right into a dataset, and create a dashboard displaying the sentiment evaluation of every evaluation.

Kendall Ruber shared her code with us, and we’ve recreated a dashboard beneath in Domo utilizing our Jupyter Pocket book integration. Our Jupyter integration permits for extra superior evaluation methods and fashions to be developed and deployed fully inside Domo with ease.

As an illustration, as a substitute of doing one thing out of the field comparable to a phrase cloud to attempt to determine what sentiment is (which might be completely different from the App Retailer Scores), we are able to get a greater thought of our customers’ sentiment by using Python’s Pure Language Processing strategies inside Domo. We run the critiques via an emotion detection mannequin, one Kendall discovered on Hugging Face, and with minimal effort we’ve an ML mannequin developed and deployed.

We can also reap the benefits of the scheduling choices inside Domo. As an illustration, this Dashboard beneath will replace day by day at 09:00 UTC. Different choices are additionally accessible, and present Domo customers will discover the choices acquainted.

Moreover, we herald information from the Google Play Retailer API via python code written in Jupyter. This places information from third celebration APIs that don’t but have a Domo Connector inside attain of your Domo setting.

Lastly, we additionally now assist sharing and collaborating on notebooks, in addition to integration with Accounts inside Domo.

Dashboard beneath, code and directions discovered on our GitHub web site.

Supply hyperlink