Articles

Blockchain users like the decentralized nature of the assets they hold. To exchange them, they rely however on centralized platforms. NFT users use platforms like OpenSea , Rarible or SuperRare . We analyse here the web traffic of five of the most popular platforms. First, OpenSea is the clear leader in terms of traffic with 88% of it. Second, Twitter provides 75% of the social traffic of these platforms. Third, the users are mostly men, nearly 70%, and 80% are less than 45 years old. Many more numbers are availble in the present article that gathers all the details you need. Hopefully it will be useful to other builders.

How many users aret there really in the NFT Space ? Blockchain gives very transparent insights on wallet activities that help us understand who transacted how much to whom. In the present article we try to understand how many are using NFTs. Doing se we explore first how many active users there are in the crypto space today. Second, we do the same for the NFT space and try to give more insight on the number of NFT users

The current Bear Market is often seen as positive for the Web3 ecosystem and especially by builders. Successful companies have been built during the last one. Some successful examples are OpenSea and Dune. The present article covers two aspects of the present market downturn. First, it gives a quantified overview of the NFT Market in terms of transactions, traders and exchanged volumes. Second, it gives some examples supporting the fact that the Bear Market is an opportunity to build on free NFT use cases.

Today I am at the NFT NYC conference, the biggest conference this year on NFTs. As the market is turning bearish, tt's an opportunity to reflect on the fundamentals of the market. In the present articles I am trying to answer the following questions : - Where can you buy NFTs ? What are the most famous Platforms ? What are their rate takes ? - What can you buy ? How many collections are there, what’s their size , what are the most famous ones ? - Who is actually buying NFTs and how many of them ?

Adopting a Data Centric approach to a Machine Learning project will probably leave you wondering what steps will lead you there. Improving the data rather than the model seems quite obvious since everyone knows that the garbage in yields garbage out. The previous article discussed the different challenges that must be overcome to reach Data Excellence. The goal was to formalize what are the data challenges a Data Scientist might encounter and what impact they can have on a Machine Learning project.

A few months ago, Andrew Ng initiated the Data Centric AI movement. He suggests that to improve the performance of a model on a given task improving the dataset will ultimately lead to better results than iterating on the model itself. While it seems obvious that better data leads to better models, improving the data can remains vague. Where some choices are sometimes made intuitively, a precise and technically rigorous methodology is required. This can lead to errors in replicability, consistency, dataset versioning, etc. I questionned few of my data scientists peers on the topic. According to them Data Science is an Art. It is clear that intuition and practice matter in the field. Nevertheless I think a lot can be learned from Data Engineers to improve the strentgh and quality of data science projects. Defining data excellence is difficult. The purpose of the present document is therefore to help the reader to understand more clearly the traps to avoid to ensure successful project outcomes.