Exercise

Sentiment analysis on formatted text

In this exercise, you'll calculate the sentiment on the customer channel of call_2.wav.

You've split the customer channel and saved it to call_2_channel_2.wav.

But from your experience with sentiment analysis, you know it can change sentence to sentence.

To calculate it sentence to sentence, you split the split using NLTK's sent_tokenize() module.

But transcribe_audio() doesn't return sentences. To try sentiment anaylsis with sentences, you've tried a paid API service to get call_2_channel_2_paid_api_text which has sentences.

Instructions 1/3

undefined XP
  • 1
    • Transcribe the audio of call_2_channel_2.wav and find the sentiment scores.
  • 2
    • Split call_2_channel_2_text into sentences and find the sentiment score of each sentence.
  • 3
    • Split call_2_channel_2_paid_api_text into sentences and score the sentiment of each.