Marketing Multi-Channel Attribution model with R (part 1: Markov chains concept)
Attribution modelling in Marketing is a new term, after digital advertising (mostly Facebook and Google) became popular. Attribution is basically the method companies or advertisers use in order to measure performance of their advertisement channels. The method tries to solves which ad channel gets the score after each purchase or conversion. The problem arised after users or customers begin to use lots of different tools and channels before they buy a product. A customer first sees the product from a Facebook Ad, Googles it, sees the ad again (Facebook Retargeting) and visits the website organically (Browser) and purchases the product. So which ad is most influential for the user to make a purchase? The work on the web page tries to solve this problem with Markov chains.
Exploratory Analysis of Mass Shooting in USA
EDA of last weeks shooting in a concert in USA. It summarizes past mass shootings with some visualization. (Using Python)
Zillow is the most famous real estate listing site the USA. It has started a competition on Kaggle, in order to increase their Zestimate accuracy. Zestimate is an automated online valuation tool which tries to predict real estate value. The work is basically summarizes competition data using EDA and visualization.
Basic Machine Learning with Cancer
ML tutorial for beginners. The work tries to predict whether a tumor is malignant or benign. After tuning parameters, he increased the accuracy from 70% to 95%. (Using Python)
Simple EDA for football players and radar chart using R and visualizations.
FBI and Justice Department Homicide datasets are analyzed. The variables of more than 22K observations are age, race, sex, ethnicity of victims and perpetrators, the relationship between the victim and perpetrator and weapon used. Kumar, the analyst, tries to answer trend of homicide over the years, most used weapon for killing, susceptable age groups and famicide.
Exploring Hillary Clinton’s Emails
Hillary Clinton’s 7,945 emails in response to a FOIA request are analyzed. Common terms and subjects are revealed.