ruchadesh/data_science_portfolio

** Data Science Projects**

These are some of the data science projects I got the opportunity to work on during my Masters course. I worked on this projects using R Studio and Python.

[Text Mining (Sentiment Analysis)]

(https://github.com/ruchadesh/Data_Science_Portfolio/tree/master/Text%20Mining%20(Sentiment%20Analysis)):

[Bank Marketing data analysis and Prediction]

(https://github.com/ruchadesh/Data_Science_Portfolio/tree/master/Bank_Marketing%20Research%20Analysis%20and%20Prediction):

• This project consists of a Machine Learning model to predict if a client will subscribe to the product, given his/her demographic and marketing campaign related information

Employee Churn Prediction (HR Analytics):

• People analytics can help to assess the effectiveness of people practices, programs, and processes. This project has helped me understand how knowledge of social and data sciences can help you make more informed, objective people decisions.

• HR Analytics includes the application of analytical processes to the Human Resources department of an organization in order to increase the employee performance and get a better return on investment.

Health Insurance MarketPlace Analysis and Prediction:

• Health Insurance is a program that covers medical expenses through private/ social or social welfare program . Health Insurance has
  started declining since 2000. Due to unemployment and rise in insurance cost, the rates of coverage have declined significantly. A rise is observed in public insurance since the pool of people with private health insurance has shrunk.

• This projectc helps the customers to select a suitable metal plan category according to their needs.

Kaggle competition- Titanic Survival:

• The wreck of the RMS Titanic was one of the worst shipwrecks in history, and is certainly the most well-known. On April 15, 1912, during her maiden voyage, the Titanic sank after colliding with an iceberg, killing 1502 out of 2224 passengers and crew. This
  sensational tragedy shocked the international community and lead to better safety regulations for ships.

• One of the reasons that the shipwreck lead to such loss of life is that were not enough lifeboats for the passengers and crew. Although there was some element of luck involved in surviving the sinking, some groups of people were more likely to survive than others, like women, children, and the upper-class.

• This contest code includes the analysis of what sorts of people were likely to survive.

Walmart Sales forecasting :

• It contains historical sales data for 45 Walmart stores located in different regions. Each store contains a number of departments, and you are tasked with predicting the department-wide sales for each store.

• In addition, Walmart runs several promotional markdown events throughout the year. These markdowns precede prominent holidays, the four largest of which are the Super Bowl, Labor Day, Thanksgiving, and Christmas. The weeks including these holidays are weighted five times higher in the evaluation than non-holiday weeks.

Credit Card Fraud Detection:

• The datasets contains credit card transactions over a two day collection period in September 2013 by European cardholders. There are a total of 284,807 transactions, of which 492 (0.172%) are fraudulent.

The dataset contains numerical variables that are the result of a principal components analysis (PCA) transformation. This transformation was applied by the original authors to maintain confidentiality of sensitive information. Additionally the dataset contains Time and Amount, which were not transformed by PCA. The Time variable contains the seconds elapsed between each transaction and the first transaction in the dataset. The Amount variable is the transaction amount, this feature can be used for example-dependant cost-senstive learning. The Class variable is the response variable and indicates whether the transaction was fraudulant.

The dataset was collected and analysed during a research collaboration of Worldline and the Machine Learning Group of Université Libre de Bruxelles (ULB) on big data mining and fraud detection.

• Model Evaluation :

The final model achieves an overall f1 score of 1.00, with 95% sensitivity (recall) and 19% precision for the positive class. That is, the model correctly identifies 95% of the fraud cases (true positives) but only 19% of the transactions predicted as fraudulent were actually fraudulent. The model catches 95% of the fraudulent cases — it could identify more cases of fraud but would then also have lower precision.

Classification report

         precision    recall  f1-score   support

    0.0       1.00      0.99      1.00    284315
    1.0       0.19      0.95      0.31       492

avg / total 1.00 0.99 1.00 284807

Cross-tabulation

Predictions 0 1 Truth
0.0 282274 2041 1.0 26 466