Soni Rani Email and Phone Number

1. Built a clustering model using K-means to identify major customer segments on demographic and psychographic-based data to optimize impact of marketingcampaigns.2. Engineered features to obtain new features such as RFM, RFMGroup, and RFMScorefor getting more details about the customers' purchasing behaviour. 3. Evaluated the optimal clusters using Silhouette score and Elbow method and leveragedthe visualization library t-SNE for multidimensional scaling to visualize and validate theinter-cluster separation and intra-cluster similarities.4. Deployed K-Means clustering and got the optimal number of clusters equal to 2 havingthe highest silhouette score of 0.39.

1. Developed a robust book recommendation framework using memory and model basedcollaborative filtering by utilising the ratings of books and various user features.2. Implemented Singular Value Decomposition based Matrix Factorization to obtainuser-item interactions and employed cosine distance to measure user-item similarities.3. Created profiles for top active users by leveraging interaction strength with therecommended items and achieved test recall@5 of 42% and recall@10 of 53%.4. Handled the cold start problem based on global and demographic-specific bookpopularity and improved the efficiency of the user recommendation engine by 33%.

1. Developed a binary classification model using algorithms such as Logistic Regression, SVC and XGBoost to predict whether a customer will default on credit card payments.2. Extraction of features from the customers datasets followed by data analysis usingStatistical and Visualization tools.3. Implemented SMOTE boosting to oversample the minority class observations andcarried out hyperparameter tuning. 4. Deployed XGBoost classifier to predict whether a customer will default and achievedhighest test accuracy score of 87% and AUC is 0.873.

1. Developed a regression model using algorithms such as Random Forest and XGBoostto predict the hourly demand of bikes and reduced the public waiting time.2. Designed the data pipeline to work on, understood the impact of features such as time of the day, weather condition, holidays and engineered important features for model.3. Imputed missing values, encoded categorical columns, handled outliers, checked multicollinearity and experimented with different models. 4. Performed hyperparameter tuning techniques such as GridSearch CV and achieved R2score of 92% using XGBoost model and reduced the public waiting time significantly.