Exploring data, understanding models, and building practical applications.
I'm Fatima Sohail, a BS Data Science student interested in machine learning, data analytics, and full-stack development.
My projects range from implementing machine learning algorithms from scratch to analyzing large datasets with Spark and building interactive applications. I enjoy understanding what happens behind a model's predictions and presenting the results clearly.
- 🧠 Machine learning: classification, clustering, feature engineering, and model evaluation.
- 📊 Data analytics: cleaning datasets, exploring patterns, and creating visualizations.
- 🔬 Research projects: medical image classification and motor fault diagnosis.
- 💻 Application development: interactive R Shiny tools and web projects.
- 🤝 Open to: internships, collaborative projects, and opportunities to apply my skills.
📚 Languages, Libraries & Tools
| Area | Technologies |
|---|---|
| Programming | Python · R · SQL · C++ · Java · JavaScript |
| Data Analysis | Pandas · NumPy · Matplotlib · Seaborn · ggplot2 · dplyr |
| Machine Learning | Scikit-learn · Classification · Clustering · PCA · Feature Engineering |
| Deep Learning Projects | EfficientNet · Graph Attention Networks · Grad-CAM |
| Big Data & Databases | Apache Spark · Hadoop · PostgreSQL · Data Warehousing |
| Web & Interactive Apps | React · HTML · CSS · Bootstrap · R Shiny |
| Parallel Computing | OpenMP · CUDA · CPU–GPU Computing |
| Development Tools | Git · GitHub · Jupyter Notebook · VS Code · Docker |
Studying motor health through electrical current signatures, with from-scratch implementations of K-NN, PCA, Logistic Regression, Naive Bayes, and SVM.
The project explores binary and multiclass classification, dimensionality reduction, and class-level performance.
Focus: Machine Learning Signal Data Model Evaluation
A research implementation combining EfficientNet-B0, a Region-Graph Attention Network, and a Fuzzy Confidence Head for five-class cervical cell classification.
Includes cross-validation, interpretability, and ablation experiments.
Focus: Deep Learning Computer Vision Research
Analysis of 2.96 million taxi-trip records using Apache Spark in local mode, covering data cleaning, transformations, SQL queries, window functions, and caching experiments.
Focus: Apache Spark SQL Big Data Analytics
Customer segmentation using RFM analysis and K-Means clustering, with preprocessing, feature scaling, and cluster evaluation.
Explores purchasing behavior to distinguish customer groups.
Focus: Customer Analytics Feature Engineering Clustering
A K-NN workload comparing serial CPU, OpenMP, CUDA, and hybrid CPU–GPU implementations on 500,000 synthetic transactions.
The controlled K = 5 experiment recorded 17.84× compute speedup and approximately 2.17× end-to-end speedup for the hybrid implementation.
Focus: C++ OpenMP CUDA Performance Analysis
📂 More Projects — Applications, Analytics & Data Systems
| Project | Description |
|---|---|
| R Shiny Data Cleaning App | CSV upload, data-cleaning controls, previews, and cleaned-data export. |
| Movie Data Visualizer | Interactive scatter plots with variable selection and customizable plot controls. |
| Pakistan E-commerce Visualization | Exploratory analysis and visualizations using Python. |
| MediScan AI | A medical scan classification web application project. |
| RescueSync | An emergency response platform with web and mobile applications. |
| Tour Management Data Warehouse | PostgreSQL data warehouse design with dimensional schemas. |
| Pizza Accessibility Analysis | Spatial accessibility analysis using QGIS. |
Language statistics describe repository contents, not proficiency levels.
Each square represents a day. Green intensity reflects contribution activity.
I'm continuing to develop my understanding of:
- Model evaluation: looking beyond accuracy to understand errors and class-level performance.
- Deep learning: exploring image classification and model interpretability.
- Data engineering: working with Spark, SQL, and data warehouse design.
- Application development: connecting analytical work with useful interfaces.
- Performance: understanding the trade-offs between CPU, GPU, and hybrid execution.
I'm interested in data science and machine learning internships, analytics projects, and web development collaborations.
If you're working on something related, I'd be happy to connect.
