My Data Science Blog

A personal blog about Data Science

AI Paradoxes: Why Human Expertise is the Ultimate Moat

Automation often fails due to fundamental mathematical and psychological contradictions. From Simpson’s Paradox in causal inference to Ramsey Theory in big data, explore why the future of Data Science relies on human "Causal Auditors" to separate signal from noise.

Posted by Julian Quernheim on February 17, 2026

Chat with my personal AI assistant!

We’re moving from single-model chatbots to AI ecosystems, where agents not only reason but also cooperate. The future of AI isn’t one assistant—it’s teams of them, working in harmony to create more natural, useful, and personal experiences.

Posted by Julian Quernheim on August 08, 2025

Recurrent Neural Networks applied to predict stock market data

How do recurrent neural networks work and how can we apply them to timeseries data to make predictions about the future?

Posted by Julian Quernheim on January 06, 2019

Fake news classifier using Python

Identifying fake news from real news is a nice use case for building a classifier model. Read more to find out how fake news are identified using python and how I built a custom web application to visualize the results during a hackathon at JetBrains Munich.

Posted by Julian Quernheim on July 16, 2017

Face Recognition With Python

Face recognition is probably one of the most popular computer vision problems. The idea is easy to grasp and it is actually not so difficult as one might imagine. Check out my article on face recognition using Python and implement your first face detection application.

Posted by Julian Quernheim on April 1, 2017

Real Time Analytics with Python and Plotly (Part 1)

Real time analytics is one of the only ways to extract the most value of continuously incoming data. But how can we implement a real-time analytics application that not only does some calculation under the hood, but also presents this information in a graphical way? Check out my real time analytics article and implement your own application with Python - Kafka and Plotly.

Posted by Julian Quernheim on March 4, 2017

Data Exploration with Shiny

Data exploration can be a time-consuming task. There are often repeating analysis steps that can be applied to every data set. Check out my data exploration app which includes data exploration (missing values, histograms, ...) as well as feature reduction possibilities.

Posted by Julian Quernheim on March 1, 2017