Use both open-source libraries and SaaS APIs to build sentiment analysis solutions to automate the process analyzing the product reviews and sort them by positive, neutral, negative. Along the way, you’ll get more familiar with preprocessing, parsing the textual information, and building predictive machine learning model.
Background: Your firm starts to pay attention to the voice of customer to improve their products. You want to help product managers get insights that will help them develop a robust product roadmap and provide customers with what they want.
Your plan is to analyze the customer reviews at Amazon to understand if your customers like and dislike your product. There are a variety of valuable resources to help you start your text analysis project. You can either build your own sentiment analysis tool using open source libraries, or directly use SaaS(Software as a Service) APIs. Open-source libraries require a lot of time and technical know-how, while SaaS tools can often be put to work right away and require little to no coding experience. As your firm is limited in budget and human resource. You want to explore both resources to help you to achieve your goal.
– You collect the most recent 20 reviews by yourself from Amazon and put them in a template review.csv.
– You manually read all 20 reviews and sort them by positive, neutral, and negative.
– You train one machine learning model using open-source libraries and one SaaS MonkeyLearn for comparison.
– Your report should generate insights into three questions: