Various techniques on dealing with multi-output classification
This post evaluates different approaches when facing a multi-output within a classification problem. A multi-output describes a combination of several output...
Recent posts
Designing products with WGANs
This post elaborates on the idea of using Generative Adversarial Networks to design new products, such as car rims. Every time a company is looking for a new...
Generative Adversial Networks and Wasserstein Addition
This blog-post elaborates on the workings of Generative Adversial Networks (GANs). Particularly we are looking at the high-level mathematics and intuition of...
Using the power of transfer learning for building a flower classifier
In this blog-post we discuss the concept of transfer-learning and show an implementation in Python, using tensorflow. Namely, we are using the pre-trained mo...
Dynamic Time Warping: Explanation and extensive testing on audio and tabular data
Explaining the concept of Dynamic-Time-Warping and testing it extensively on sound and tabular data.
DengAI Challenge Second Attempt: Forecasting - Ensemble Model
Using a combination of regression, classification and smoother for our second take on the DengAI challenge.
DengAI Challenge Second Attempt: Building a scikit-learn pipeline
The workings of the underlying prediction pipeline
Classifier Evaluation Methods - A Hands-On Explanation
Accuracy/ Recall/ Precision/ Confusion Matrix/ ROC Curve/ AUC
How the People Really Voted
Why geographically correct maps show elections results inaccurately
Human vs. Machine — Reinforcement Learning in the Context of Snake
This blogpost elaborates on how to implement a reinforcement algorithm, which not only masters the game “Snake”, it even outperforms any human in a game with...
United for 30 Years — Catching up to West Germany
Visualizing 30 years of economic data between East and West Germany
DengAI: Predicting Disease Spread — STL Forecasting/ ARIMA/ Box-Jenkins
Using the STL Forecasting Method with an ARIMA model, which is parameterized through the Box-Jenkins Method.
How to create Country Heatmaps in R
One of the most powerful visualization tools for regional Panel data there is.
DengAI: Predicting Disease Spread - Imputation and Stationarity Problems
Github Repository
Polarization in the Senate
Polarization is increasing. How does this play out in the Senate?
How Polarized Is America Really?
Evidence for increasing polarization from the voting record of the United States Congress
Markowitz Portfolio Theory - An empirical explanation
Building and simulating different stock portfolio and assessing risk and return with Python and R Shiny Link to App
Empirical Evidence to common Stock Portfolio Questions
An intuitive and replicable explanation to some of the most important questions of portfolio building
Signal Processing - Engine Sound Detection
Detecting the engine-brand from a Formula 1 car by its sound using Neural Networks
Using Python to do Your Homework
Applying OpenCV and Tesseract to do your math-homework
Richter’s Predictor- Data Challenge from DrivenData
Scoring in the top one percent in the Richter’s Predictor: Modeling Earthquake Damage on DrivenData.
Multi-label Genre Predictor with Interactive Dashboard
An intuitive explanation of how to deal with a multi-label classification problem using ImdB movie-data.
And the Oscar goes to…
As an economist it is pretty frustrating to hear the news anchor saying that the cast of the new James Bond movies did it again. “The highest gross revenue e...
Conclusion - Predicting Real Estate Prices with Features Scraped from the Web
What we learned predicting real estate prices with data scraped from the web.
NLP with Real Estate Advertisements - Part 2
Applying our NLP feature vector to the Gradient Boosting model.
NLP with Real Estate Advertisements - Part 1
Analyzing and exploring real estate advertisement descriptions using NLP pre-processing and visualization techniques.
Predicting Real Estate Prices
Machine Learning for predicting Spanish real estate prices.
Feature Creation for Real Estate Prices
Feature Creation for Real Estate Price Prediction This post elaborates on feature engineering for Spanish real estate price prediction.
Clustering Real Estate Data
Using unsupervised learning techniques to create features for supervised price prediction.
Outlier detection in Real Estate Data
Preparing the data for feature extraction and modeling In the prior post, we outlined how we scraped real estate data for 42,367 properties in Spain. Before ...
Webscraping Real Estate Market Data
Using BeautifulSoup and Python to scrape real estate data from the web
Introduction - End-to-End Machine Learning for Real Estate Price Prediction
Machine learning is an extremely powerful tool, applicable to an astounding breadth of use cases. Today, almost any question imaginable can be the starting p...