In the past year, we launched multiple AI-based products. With a strong community of more than 25,000 registered users for our APIs and plugins, we are evolving every day. A persistent need that we uncovered from engaging with our community is – how can I make sense of the 1000s of customer feedback data quickly?
Between then and now, we were ideating and experimenting with a SaaS tool that can automate the task of analyzing customer responses. We wanted a solution that works in the world’s favorite format – EXCEL!
We started building this product in the month of July 2018 and after multiple iterations, we are pleased to announce our Automated Survey Analysis tool – SmartReader, which can quickly analyze large sets of open-ended verbatim comments and find actionable insights that can further your business.
You can access SmartReader from here.
In this post, we will show how to use ParallelDots SmartReader to quickly analyze a large corpus of open-ended verbatim comments and make strong, data-backed decisions.
Getting Started with ParallelDots SmartReader
To get started, Sign Up for a free ParallelDots account from here (if you already have an account then please login from here). Upon filling the form on our website and after verifying your email, you can log in directly to your SmartReader dashboard. Please store account credentials in a secure location as unauthorized access to your account will result in early exhaustion of your quota. In case you have lost your password or if you feel someone else might have access to it, please immediately change your password.
Using the SmartReader tool – A step-by-step guide
Once you will log in to your dashboard you need to follow the below steps to start building your first model for FREE!
- The first step is to click on create your name your project or ‘Model’.
- The second step is to upload a CSV file containing the data which you want to classify.
- Next, you need to select the column name which contains the data to be classified from a drop down menu.
- Finally, our algorithm will be trained on your data and it may take around 10 – 15 minutes for this step to complete. You can close the tab or carry on with other tasks.
- Once everything is trained properly, you will get a list of topics fetched from your data. You can click on any of these topics to get the keywords recommendations which define those topics (Very critical step, please review the recommended topics and their associated keywords and add/delete them if necessary).
- Define your custom categories and tags to structure your text data now. Test the results on a sample from your data to check the quality of the classification, and revise the tags until the results are as expected. There will be an input box at the bottom where you can enter text and check the quality of the classification task. Remember to provide as many sub-topics as you can!
- Once you are happy with the topics and keyword combinations, you can download the results which will have sentiment, emotion, and keywords apart from categorizing your verbatims into the categories. API access will be released soon.
In a single downloaded excel sheet, you will be able to get Sentiment, Emotion, and Keywords and binary score to each category of the review.