Blog

When business stakeholders ask data experts seemingly simple questions, they often expect a quick and straightforward answer. On the surface, it seems like a piece of cake. But in the messy reality of data, what appears to be a simple question can quickly turn into a multi-layered onion of a problem - each layer revealing increasing complexity and ambiguity. Example from the insurance domain: Let’s take a practical example from the insurance industry. A sales executive asks: ...

In recent years, in light of the growing attention given to climate change and environmental concerns, insurance companies worldwide have been increasingly focused on ESG (Environmental, Social, and Governance) objectives. Indeed, according to a survey conducted by PricewaterhouseCoopers in 2022, 85% of global insurers believe that ESG will impact all functions of their business in the years to come. Bitcoin on the other hand, is often denounced to be “terrible for the environment” and a “contributor to climate degradation”. Media outlets oftentimes quote statistics such as: ...

Insurance is a market-based risk management tool that is fundamentally based on sharing and diversifying risks among a community of policyholders. It creates a safety net against unpredictable events, protecting individuals and societies from potentially hazardous financial losses. This invaluable system not only provides a sense of security to its members but also fosters innovation and exploration, encouraging humans to embrace calculated risks. Thus, insurance makes our world more resilient, fostering innovation and stability. ...

In a recent article and also on a Twitter thread, American entrepreneur, investor, and influencer Anthony Pompliano has called Bitcoin ‘the largest insurance company in the world’: Is bitcoin the largest insurance company in the world? This was the question posed by two investors at breakfast yesterday. In this piece, I unpack why they may be correct in their analysis.https://t.co/nSS4xDnuou — Anthony Pompliano 🌪 (@APompliano) May 19, 2023 His argument in the article is quite straight-forward: ...

The first Bitcoin was mined in 2009 – only 14 years ago. In these 14 years, Bitcoin has developed from a curious experiment ran by a small group of cypherpunks into a global phenomenon. Its journey from obscurity to mainstream recognition has been nothing short of remarkable, as has been reflected not only by its price, but by the growing computing power securing this global network. Bitcoin is on its way to disrupt traditional financial systems, challenging the centralized control of money and empowering individuals with financial sovereignty. ...

Recently, I read a thread on Twitter about several Machine Learning papers that contained severe cases of data leakage. The authors of the papers seemed unaware of this phenomenon and therefore trained models that performed exceptionally well. Unfortunately, this was mainly due to data leakage. Not many beginners are aware of this problem and in my opinion, not many courses emphasize this issue early enough. Therefore, I would like to tell you all the things you need to know about data leakage and some ways to prevent it in this post. ...

Counterfeit money is a real problem both for individuals and for businesses. Counterfeiters constantly find new ways and techniques to produce fake banknotes, that are essentially indistinguishable from real money. At least for the human eye! Identifying forged banknotes is a typical example of a binary classification task in Machine Learning. If we have enough data of both real and forged banknotes, we can use this data to train a model that can classify new banknotes as either real or fake. ...

Linear and Logistic regression are among the most elementary algorithms for supervised learning. Supervised Learning describes the situation where we deal with labelled data, which means that we have labelled inputs and a target variable. Despite the fact that both have the word “regression” in their name, only one of them is typically being used for solving regression problems! Let’s see how they work! Linear Regression Linear regression is possibly the easiest, most intuitive way of making a quantitative prediction. The relationship between an independent and a dependent variable is assumed to be linear, meaning that the dependent variable can be predicted using a linear function of the independent variable. For example: ...

After reading this article from Pratap Vardhan with great interest, I wanted to build my own version of a Bar Chart Race that is smoother and a bit more beautiful. The biggest improvement is the interpolation (or augmentation) of the available data points in order to make the animation smoother. Here is the Bar Chart Race we are going to build in this article: For the purpose of this demonstration, we are going to use a GDP per capita forecast dataset provided by the OECD. You can find the original dataset here. ...

k-Nearest Neighbors, or k-NN as I am going to call it from now on, is one of the easiest algorithms to solve classification tasks. It can be used for regression problems as well, but I am going to focus on the more common use case of classification in this post. In a nutshell, k-NN will assign a new data point to the class that the majority of its k neighbours in the training set belong to. Let’s use another coffee-related example to see how that works. ...