The following section explains the feedback obtained from analyzing the text data.
- It is useful to collect data across multiple rows and group the results by a dimension. In this case, the dimension is label. A df.groupby() function is used to measure the count of 100 therapy transactions online distributed by label. We can see that there is a 65:35 distribution of do_not_escalate to escalate as seen in the following screenshot:
- A new column, word_count, is created to calculate how many words are used in each of the 100 transactions between the chatbot and the online visitor. The newly created column, word_count, can be ...