At Haptik, we build new chatbots on a daily basis. These chatbots generate data, lots of it. Data including, but not limited to, what messages were exchanged, what data elements were used by the bot to respond appropriately and what problems were detected along the way. This amounts to huge volumes of data generated daily. Clearly, there is a need to interpret that data into meaningful, insightful analytics
This necessity led to the development of our very own data-driven analytics tool, an Analytics Dashboard. In this blog, we will talk about the capabilities of our Analytics Dashboard and what went into developing it.
What is a Data-Driven Approach?
Our Analytics Dashboard is a data-driven tool, one focused on information that shows its users the behaviour and patterns created by all their end users. By observing the data constantly, it becomes possible to gain real insights of what works and what does not, what people use and what is ignored, eventually helping us in making really important strategic business decisions.
This dashboard is currently our source of truth, it is the go-to place to analyze and observe exactly what’s transpiring between our end users and all our chatbots.
Analytics at a Glance
We have a lot of data generated by various parts of our chatbot environment, all pushed to an AWS Kinesis Stream; data from this stream is then pushed to AWS Elasticsearch using Logstash.
The goal of this tool is to make sense of all the data generated by our chatbots across various businesses and clients. Businesses include but are not limited to bots created for our own Haptik App and for external clients. All the data visible inside the tool is filterable based on selected businesses, bots and over a time range.
For Example, we have an entertainment bot, which chats with users, sends across jokes, horoscopes basically spreading happiness. This bot falls under the business of the Haptik app. Now we can see which components of the bot are used most frequently, what new detected keywords are causing the bot to break, what’s working and what’s not for a given time range.
The main technologies used to create this tool are as follows:
If you would like to know why we choose ReactJs for our tools, please refer to the blog here.
What Analytics Do We Provide?
Our overall analytics is divided into the following major components:
What Did We Learn From This?
We will continue to experiment with more technologies out there and make this tool even better as a benchmark for Analytics Tools. Do let us know if you have any feedback in the comment section below; we will be happy to know.
Haptik is hiring. Do visit our careers section and get in touch with us at firstname.lastname@example.org.
This content was originally published here.