Carmel Dansou

Data-driven professional with a background in computational physics and expertise in analytical modeling and cross-functional collaboration. Proficient in Python, Numpy, and quantitative analysis, demonstrating the ability to derive actionable insights from complex datasets. Experienced in advanced modeling techniques, eager to contribute to real-life interdisciplinary projects involving data sourcing and modeling for data-informed decisions. Seeking a transition into a new field where data drives practical decision-making, aiming to leverage transferable skills within a dynamic and innovative setting.

• Google Advanced Data Analytics Certification
• Bayesian Methods for Machine Learning
• TensorFlow for AI and Machine Learning

• Predict airline customers satisfaction using XGBoost: Building a XGBoost classification model to predict whether an airline customer is satisfied or dissatisfied. Data splitting using train-test-split function to train and test data subsets. Performing hyperparameter tuning using GridSearchCV function. Representing best parameters and scores of the XGBoost model. Model achieves F1 score of 0.937, which suggests very strong predictive power in this model.
• Classifying TikTok videos using machine learning: Build Random Forest, XGBoost classifiers to predict whether a TikTok video presents a 'claim' or presents an 'opinion'. Perform exploratory data analysis, data splitting in train, validation and test subsets. Fit models and tune hyperparameters on the training set. Perform final model selection on the validation set. Assess the best model's performance on the test set.
• Build a multiple linear regression model to predict taxi fares: Exploratory data analysis: identifying outlier, missing values, data visualization. Feature engineering and encoding of categorical variables, collinearity analysis. Separate clean data into train and test data subsets, standardize the data. Model building and evaluation using Mean Absolute Error, Mean Squared Error, and the Root Mean Squared Error. Model achieves R2 = 0.84, RMSE = 4.2 on train set and R2 = 0.86, RMSE = 3.78 on test set suggesting model has little bias.

• Collaborative team member with strong interpersonal communication skills.
• Highly adaptable and eager to contribute to practical data driven decision making projects.
• Proven track record of self-directed learning and mastering complex technical domains.