Automated-Customer-Review-Analytics

[gminova]

Name of proposer Gigi @gminova

Describe the problem you'd like to solve Manually reading and analyzing reviews can be a daunting, time-consuming and expensive task. Creating an algorithm (Automated-Customer-Review-Analytics) which analyzes customer reviews and is able to distinguish between a positive, a neutral and a negative review can streamline the process of cathegorizing reviews. To achieve that, the team will look into natural language processing (NLP), text mining and other forms of artificial intelligence (AI). Customer reviews will consist of 3 main pieces of information:

1. Overall star rating (1-5) :star::star::star::star::star:
2. A title / headline of review
3. Text with the customer review / feedback

Who are the stakeholders?

• Who would this benefit? E-commerce businesses who would like to look into ways of improving their customer experience and improving their services.
• Who would be affected by it? Customers, product researchers, business stakeholders, brands and online commerce sites.

What is currently being done to solve this problem?

Examples of current innovative solutions looking into this problem:

1. https://www.revuze.it/
2. https://www.lexalytics.com/
3. http://sentigeek.com/

[samjam48]

Whoa, this would be a cool project. Some cool NLP stuff out there that might let you do this quickly. but could be a big thing to take on.

This site is also cool in terms of working out spammy reviews and such to give a more even calculation.

https://reviewmeta.com/