DIGVIJAY PARMAR

Optimized the workforce for different projects based on defined timelines using MS Excel & Visualisation. Managed financial resources through the use of MS Excel to ensure efficient and effective financial management. Facilitated client engagements to develop & coordinate the end-to-end bid process with senior leadership. Used 'ServiceNow' incident event log data from UCI Machine Learning Repository to predict time it will take for a ticket to be resolved. Missing values in the categorical columns were handled using mode imputation, weighted mode imputation and factorize (for non-mandatory columns). New features were created using datetime components, target encoding for highly cardinal columns, followed by heatmaps to check for correlation with the Target Variable. Non-Parametric tests Kruskal-Wallis & MannWhitneyU to statistically verify results from EDA. Various Regressors including Linear Regression, Decision Tree, KNN, Random Forest, XGBoost were used (5 fold cross validation) with the maximum R-squared (0.94) and minimum MAE (2.4 days) on XGBoost. Linear Regression (0.81 and 4.11 days) HealthCare data from Analytics Vidhya was used to predict the chances of a favourable outcome (getting a health score or visiting a stall). Missing values were imputed with median (numerical columns) and factorize (categorical columns). Feature Engineering including duration of camps, target variable, datetime components were created. Hypothesis testing using Chi Square test to statistically verify results from EDA. Various Classifiers including Logistic Regression, Decision Tree, Random Forest, LightGBM, XGBoost, CatBoost were used (5 fold cross validation) with maximum ROC-AUC score (0.727) with LightGBM followed by XGBoost (0.71). Used genes data which was anonymised to discover groups of correlated genes that are potentially associated with a disease or condition. Various Clustering Models including KMeans Clustering, Agglomerative Clustering and DBSCAN were implemented and