Data Visualization : Bringing Customer Engagement in Solving ProblemsSudhanwa Rajurkar
In today’s Big Data problems and Data Science solutions, Data Visualization is becoming more important than ever. When the data is large, you need an efficient way of communicating the results. This is what Data visualization is all about. It makes data more natural for the human mind to comprehend.
Consider a scenario, where a Data Science team is working on Customer tickets, to come up with ML models for automatically tagging these tickets with relevant tags.
The team has trained a few models on some labeled data and has predicted results for 50000 un-tagged tickets.
They now want to send this information to the customer to validate the tags and identify the models which are more effective.
How should they send these 50000 results to the customer to validate the predicted tags?
One simple answer is to zip these 50,000 results, send them to customers, let them review and validate the tags. After all, these are intermediate results.
Based on their answers, we will identify the effective model and accurate, important tags.
Do you think the customer will have the time and motivation to go through 50K results?
Even if they do it, the DS team would need answers recorded in a particular format, and if the results are well-formatted, these need to be further processed to arrive at information that the DS team is seeking from this review
50K test data is a very modest size in the world of Data Science. We can’t charge our customers to do this cumbersome work, which has resulted from our inefficiencies.
The DS team may have to keep waiting for the customer for the results. No one knows when it will arrive.
It’s the DS team’s job to present these results in an efficient manner so that our customers can easily comprehend consume, identify patterns, outliers that they are most interested in.
The visualization should be such that our customers should be excited to look at the results and give us our answers in a few minutes.
Data should give them joy, not grief! That is a sign of good data visualization.
Data Scientists who understand the importance of Data visualizations and use them in their development process as well are a dream to work with.
Let me leave you with the problem, and allow you to think of some data visualizations for the same.
What Data Visualization solutions do you propose?