✽ Pamela Martinez Portfolium

Data Science

In this segment, discover my journey through the multifaceted world of data science. This section showcases my expertise in various aspects of data science, including machine learning, statistical analysis, data mining, and predictive modeling. The projects featured here solve real-world problems and highlight my proficiency in using Python, Jupyter and Azure Notebooks, SQL, TensorFlow, as well as Scikit-learn for machine learning algorithms, Pandas for data manipulation, Matplotlib and Seaborn for data visualization, Keras for deep learning, and Hadoop for managing large datasets. From crafting advanced forecasting models to executing detailed sentiment analysis, these projects underscore my dedication to using a wide array of data science technologies for strategic decision-making and innovative problem-solving.

Back to Home

Data Science

In this segment, discover my journey through the multifaceted world of data science. This section showcases my expertise in various aspects of data science, including machine learning, statistical analysis, data mining, and predictive modeling. The projects featured here solve real-world problems and highlight my proficiency in using Python, Jupyter and Azure Notebooks, SQL, TensorFlow, as well as Scikit-learn for machine learning algorithms, Pandas for data manipulation, Matplotlib and Seaborn for data visualization, Keras for deep learning, and Hadoop for managing large datasets. From crafting advanced forecasting models to executing detailed sentiment analysis, these projects underscore my dedication to using a wide array of data science technologies for strategic decision-making and innovative problem-solving.

Back to Home

Clinical Triage NLP: TF-IDF
Naive Bayes vs Zero-Shot

Classifies clinical notes into mild / moderate / severe. Baseline TF-IDF + Multinomial NB versus a zero-shot transformer. Includes confusion matrix and group-wise F1 (gender & age bands) as a quick fairness check.

Link to Github

Credit Scoring on Tabular Data:
DNN vs ResNet + SHAP

Credit scoring with SMOTE for imbalance, comparison of a Dense Neural Net and a Tabular ResNet. Evaluates accuracy/ROC-AUC and business cost (FP/FN). SHAP explains feature impact on decisions.

Link to Github

CIFAR-10 CNN: Regularization & Training Dynamics

Image classifier with BatchNorm, L2, Dropout, data augmentation and callbacks (EarlyStopping, ReduceLROnPlateau, ModelCheckpoint). Reports accuracy/loss curves, learning-rate schedule, confusion matrix, and an error gallery.

Link to Github

Create a free website with Framer, the website builder loved by startups, designers and agencies.