WE CAME, WE SAW, WE SOLVED
We used three internal sources and 8 external social media sources to consolidate data at various hierarchical levels and categories. Then our data science team created a machine learning model to understand emotions in textual data, converting texts and emojis into tokens, identifying emotions from those tokens. The team also created an action plan, action templates, and reports based on the findings.
The model continues to finetune as we enter second year of implementation.