Projects
Project Title:
Titanic Survival Prediction
Objective:
To predict the survival of Titanic passengers using machine learning models based on features like age, gender, passenger class, and other relevant attributes.
Problem Definition
-
Predict survival outcomes for Titanic passengers using historical data.
-
Classification problem: Binary outcome (Survived = 1, Did not survive = 0).
Dataset Overview
-
Source: Titanic dataset from Kaggle.
-
Features:
-
Passenger Information: Name, Age, Sex, Ticket, Cabin, etc.
-
Travel Information: Pclass, Embarked, Fare, etc.
-
Target Variable: Survival (0 or 1)
-
Achievement:
-
Created predictive models that increased Titanic survival forecasting accuracy by 15% -
Achieved 92% accuracy (R² = 0.92) in predicting sales from advertising spends across TV,
Radio, and Newspaper channels, using R², MAE, MSE, RMSE and MAPE for model evaluation.
-
Delivered insights through clear visualizations to support informed decision-making.
Project Title:
Building a Real-Time Object Detection System
● Objective:
Implement an object detection model that can detect multiple objects in real-time from a live camera feed.
● Dataset:
Use a pre-trained dataset like COCO or Pascal VOC, or create a small custom dataset.
● Project Steps:
○ Data Preparation:
If creating a custom dataset, label images using tools like LabelImg.
○ Model Selection:
Use a pre-trained object detection model like YOLO, SSD, or Faster R-CNN.
○ Model Training/Fine-Tuning:
Fine-tune the model on the custom dataset if necessary.
○ Evaluation:
Use metrics like Mean Average Precision (mAP) and IoU (Intersection over Union) to evaluate performance.
○ Deployment:
Deploy the object detection model in a live-streaming app using OpenCV and Flask.
● Stretch Goals:
Add custom classes, implement multi-object tracking, or use TensorRT to optimize for faster inference
Project Title:
Sales Prediction using Python with Machine Learning and Jupyter Notebook
Objective:
Forecast future sales based on various factors such as advertising expenditure and target audience segmentation.
Achievement
Achieved an R² of 0.92, with a Mean Absolute Error (MAE) of 1.16 and Mean Absolute Percentage Error (MAPE) of 8.88%, providing accurate
advertising sales forecasts across Lasso and Ridge models.
Project Title:
Movie Rating Prediction with Python
Objective:
-
Objective: Build a model to predict movie ratings based on features like genre, director, and actors.
-
Tasks Completed: Task 2
-
Dataset: IMDB India Movies Dataset
-
Welcome to my website project page. Kindly visit my GitHub page for more projects; this is just the tip of the iceberg.
https://github.com/Tomtwiny121
Loops-And-Functions-Code-Challenge
Challenge Overview
In this coding challenge, we will apply our Python programming skills to automate more complex movement of our farm equipment. We will use functions to make our code more modular, making it simpler to read, modify, and extend.
Visualizations
CRIME-RELATED DATA VISUALIZATION IN POWERBI AND STORY TELLING
This project is an excerpt that focuses on the crime-related data in five provinces as indicated in the map. This data shows how females, males, and children are exposed to crimes in different water source locations with regards to various towns
Road Accident Dashboard Project
Excel Visualization of a Road Accident dynamic Dashboard showing Total, Fatal, Serious, Slight casualties; casualties by vehicle & and road type, location, road surfaces, & and light condition. It includes links for the database (analysis sheet), dashboard designs, datasheet, KPIs, and urban & and rural areas affected
Sales Data Analysis Report-Project
Objective:
Objective: To examine sales data and identify trends, best-selling products, and essential revenue metrics for strategic decision-making.
Undertaking: Undertook a thorough analysis of a significant sales dataset.
Insights: Explored sales patterns, identified top-performing products, and calculated crucial revenue metrics.
Visual Representation: Developed visually compelling representations to effectively communicate the findings.
Project Outcome: Offered valuable insights to support well-informed strategic business decisions.
HR-Employee Attrition-Showing the Demographics Dashboard
This project analyzes employee attrition data and builds a dashboard to visualize demographic trends and attrition patterns. HR records are used to calculate attrition rates by attributes like gender, age, location and tenure. Predictive models identify key risk factors. An interactive visualization dashboard is developed allowing filtering and comparisons of attrition percentages across demographic segments. Insights help HR focus retention efforts by pinpointing high attrition groups. The goal is to reduce unwanted turnover through informed analysis and data-driven decision making.
