Hover over and click on the highlighted countries to see the high level project descriptions and graphs:
Worldwide COVID-19 Analytics & Forecasting
This dashboard contains a comprehensive overview of the progression of Coronavirus COVID-19 spread. Analytics, and forecasting are available for any continent and country. The COVID-19 data is updated once a day around 05:00 (UTC) - 3pm in Australia and 7am in France.
This is a project that I did in early March 2020:
Being French living in Australia and having friends and loved ones in France but also all around the world, I was curious about the global spread of COVID-19.
I was particularly interested in seeing the cumulative number of cases for each country over time, as well as new cases emerging by day. So I decided to develop a web-based app/dashboard that gives me all this information in one place.
I hope you find it informative.
Anomaly detection
Anomaly/Outliers detection using Z-score for time series on New York city taxi dataset and on 2d data.
Time series Forecasting
Time series Forecasting using XGBoost and GridSearch for parameters optimisation on New York city taxi dataset.
Predicting price of diamonds
Predicting price of diamonds using Decision Tree and GridSearch for parameters optimisation.
Mall Customers Segmentation
Mall Customers Analytics to identify potential clusters that can drive marketing strategies.
Customer Insights Analytics & RFM Segmentation
Recency, Frequency and Monetary Value Analysis (RFM) segmentation is a great method to identify groups of customers for special treatment. While there are countless ways to perform segmentation, RFM analysis is popular for three reasons: It utilizes objective, numerical scales that yield a concise and informative high-level depiction of customers. It is simple – marketers can use it effectively without the need for data scientists or sophisticated software. It is intuitive – the output of this segmentation method is easy to understand and interpret.
Product Recommendations
Recommending products Using Singular Value Decomposition (SVD) by searching a large group of people and finding a smaller set of users with tastes similar to a particular user.
SVD is a classical Collaborative Filtering method from linear algebra popular for developing recommender systems.
Naive Bayes Classifier from scratch
Naive Bayes Classifier from scratch on the iris dataset.
In statistics, naive Bayes classifiers are a family of simple “probabilistic classifiers” based on applying Bayes’ theorem with strong (naïve) independence assumptions between the features. They are among the simplest Bayesian network models, but coupled with kernel density estimation, they can achieve higher accuracy levels.
Logistic Regression from scratch
Logistic Regression from scratch on the diabetes dataset with a comparison with Tensorflow.
SQL tutorial
A friend asked me for a small SQL demonstration and I decided to make it as a tutorial with multiple examples.
