In this blog I’m going to give a brief overview of how we can achieve sentiment analysis.
From Data Services 4.0 the feature Text Data Processing was made part of Data Services, this allows us to extract keywords from unstructured text. These keywords are extracted based on a dictionary. So for sentiment analysis we are interested in pulling out parts of the text and determining if their is positive or negative remarks.
I have created a table with positive and negative sentiments. I have underlined the sentiments in red.
I then built a simple data flow to read the text data and store in another table with extra information. Below you can see lines highlighted related to sentiment. As seen in the below picture not only can we determine the sentiment but can classify it as a strong or weak sentiment.
We can then create reports on the data to achieve meaningful sentiment analysis.
Example of reports created in Web Intelligence.
Here is another example created in Web Intelligence where we narrow down sentiments by city.
For more information follow me on twitter @louisdegouveia