Dipankar Mazumdar

I am a Data Science professional with experience in advocating for analytics/data products. Currently I work as a Technical Advocate at Dremio, where I focus on Dremio's datalake platform along with open-source projects such as Apache Arrow, Apache Iceberg, Arrow Flight & Project Nessie. Previously, I led the Technical Evangelism team at Qlik R&D working on key developer-strategies and educating the worldwide Qlik developer community in the areas of Data Visualization & Machine Learning.

I am also a Visual Analytics researcher and my research is focused towards developing visualization techniques to understand Machine Learning models as well as multivariate data sets. Some of the tools & programming languages that I use in my work and research are: Python, SQL, NodeJS, D3.js, Qlik Sense, Plotly, Amazon SageMaker.

I completed my MSc in Computer Science at Dalhousie University, where I wrote my thesis on "Interpretability of Random Forest models"

dp976894@dal.ca / GitHub / LinkedIn / Medium
Research
Random Forest Similarity Maps
MDPI, Special Issue 10th Anniversary of Electronics: Recent Advances in Computer Science & Engineering
Dipankar Mazumdar, Fernando V.Paulovich, Mário Popolin Neto
[Paper] [Live Demo]
Projects
Visual Text Analytics using Qlik & Word Embedding technique(Word2Vec)
Dipankar Mazumdar

A full-stack Visual Text Analytics app developed using Qlik Sense OSS - Nebula.js and Picasso.js. The application leverages techniques such as Word Embeddings(Word2Vec) to derive textual insights from a Cannabis Effects dataset.

[GitHub]
Analyzing customer churn using Qlik Sense & SageMaker
Dipankar Mazumdar

This work shows how the customer churn problem can be effectively analyzed using Qlik Sense by integrating it with SageMaker. This also lays the foundation for integrating Qlik with other AutoML platforms.

[GitHub]
Multivariate data representation using RadViz
Dipankar Mazumdar

Focus of this work is to effectively visualize high-dimensional multivariate dataset in a novel radial visualization. Experiments were carried out using the Spotify's song features dataset.

[GitHub] [Technical Report]
Latest posts
Interpreting Machine Learning models using Visual Analytics
[Medium]
Building a Visual text analytics app using Qlik and Machine Learning techniques in NodeJS
[Medium]
Building your first Qlik Sense ChatBot using Natural Language API
[Medium]
Community

I am a community person and love sharing my work in my socials, Qlik Community blog and my hobby blog