App idea: Customer reviews analysis

New Member
1 0 0


We are small software company with expertise in machine learning and natural language processing that is interested in application of such technology in e-commerce domain to help sellers boost their productivity. At this moment we have an algorithm that can analyze any kind of text and extract some useful information from it:

  • Aspect - mention of product feature, characteristic, part, etc.
  • Description - phrase that describes aspect related to it in some way. For example, what user feels about this product, how well it behaves.
  • Sentiment - the indicator of the emotion that user wants to express about aspect-description pair. The combination of aspect, its description and sentiment called “opinion”.

Here is an example of opinions extracted from Amazon reviews:

At this moment we have this kind of applications in mind:

  • Beloved features widget: Show what features customers like in your products in concise form.  For each product in catalog, our algorithm extracts user opinions from reviews and selects most useful and readable to convince doubting buyers that product is really good and matches their needs. Also, we can personalise this widget and show features that relevant to particular user.
  • YouTube reviews widget: Find and use video reviews from opinion makers as powerful source of trust to your products. We analyse content of video and extract only valuable parts of it to build short summary what reviewer do like about your product.
  • Social proof popups: All extracted opinions can be shown in popup messages on store.

We want to be sure that this kind of technology  really can be useful for your daily duties to invest resources in Shopify integration. If you find such technology interesting, feel free to leave any kind of feedback (positive and negative).

We are open to suggestions and will be happy to answer your questions!