José Daniel

Skills & Specialties 🔥

#LifelongLearner #Breezy #TeamWork #Listener #CommunicatingTechnicalInfo #BusinessAcumen #Python #SQL #MachineLearning #ComputationalThinking #EngineeringThinking

Data Analytics

■ Strong SQL knowledge.
■ Advanced Python (Pandas, NumPy, Seaborn, Dash, Plotly, Flask).
■ Process Mining (Celonis).
■ Power BI | Tableau | Excel.

Machine Learning

■ Solid statistical foundations.
■ Regression & Classification: gradient descent, decision trees, SVM, XGBoost...
■ Unsupervised & semi-supervised learning: clustering, dimensionality reduction (PCA, t-SNE), label propagation...
■ Deep learning: CNNs, ResNets.
■ Libraries: Scikit-learn, Yellowbrick, Statsmodels, PyTorch, Tensorboard, Tensorflow, Auto-SkLearn, TPOT.

Web Development

■ Front-end (HTML, CSS, Bootstrap5).
■ Back-end (Python, Flask).
■ Deployment (GitHub, Docker, AWS).

Portfolio 💥

ML - Supervised Learning

■ Classifier for Marketing and Inventory data at Walmart's lab: User-journey data to classify purchasing customers from non-purchasing customers.
■ Random Forest Classifier with weight balancing: Classification to predict qualification for loan in future.
■ SVM for classification: SVM intuition + Face recognition.
■ Comparing regression algorithms: Car sales predictions in California.
■ Polynomial regression: Regression on COVID-19 data for Chicago to predict the number of infections.
■ Multivariate Linear Regression: Predicting House Price from Size and Number of Bedrooms.

ML - Unsupervised & Semi-supervised Learning

■ Categorizing new products at Target's cosmetics: AutoML vs. LabelPropagation vs. ZSL.
■ Label Propagation: Label predictions for unlabeled digit images.
■ Zero Shot Learning (ZSL): Caltech UCSD Birds data set (788 images from 200 different types of birds annotated with 312 attributes).
■ Ecommerce data for cosmetics retail: Dimensionality reduction (t-SNE in Tensorboard) and K-means clustering.
■ k-Means Clustering: Techniques and fundamentals.
■ Anomaly Detection: PyOD (Python Outlier Detector).

Data Analytics & AI, City of Edmonton

Project: Improving business workflow and customer experience by forcasting construction permitting timelines based on resource avaliability and system queues.

Development: Web Multipage Dashboard developed using SQL, Python (Dash, Flask, Plotly), and ML prediction models. Deployed using Docker and AWS LightSail.

Process Mining, City of Edmonton

Project: Using process mining mining to better understand business workflows, critical paths, bottlenecks, and further insights about the internal projects and operations.

Development: Dashboards created in the Celonis Software using thousands of event logs with information about internal business processes.

Thesis Protoype - Pt1

Project: This research proposes a BIM-based Generative Design approach to optimize the sheathing layout arrangement in a prefabricated construction environment. This approach can allow practitioners to reduce the overall waste for given designs and evaluate various layout alternatives. The framework includes a decision-support module that considers environmental, cost, and aesthetic aspects to identify the optimal layout.

Development: The proposed model was developed and implemented in Python using OOP, optimization algorithms, generative design techniques, and BIM 3D models.

Thesis Protoype - Pt2

Project: In this research, a two-story wood residential house (53 wood frames) is used as a case study to demonstrate the framework’s practical applicability. After implementation, three “best” design alternatives were found according to the decision aspects. The design improvements achieved were 37.5%, 7%, and 54% for the environmental, cost, and aesthetic factors, respectively.

Development: The script was developed using visual programming in Dynamo to visualize final Python results (optimized solutions) in the BIM 3D environment.

Read full research paper in the Journal of Advances in Civil Engineering.