NLP Web-Based App for Depression Detection

Objective: Developed a natural language processing (NLP) model that classifies Reddit posts as either depressive or non-depressive.

Technologies Used: Python (NLTK, Scikit-learn), Data Visualization (Matplotlib, Seaborn), HTML/CSS, Streamlit

Features:

Input: User submits a Reddit post or text.

Output: Displays whether the text is classified as depressive or non-depressive with a confidence score.

Deployed as a web app for real-time use.

View Live Demo

Outcome: Achieved 90% accuracy, showcasing linguistic markers of depressive tendencies.

View Project

Sentiment Analysis of Social Media Posts

Objective: Performed sentiment analysis on tweets related to Apple and Google products, classifying user sentiments into positive, negative, or neutral categories using Natural Language Processing (NLP) techniques. The goal is to assess public perception of these brands based on social media data.

Technologies Used:

Programming Languages: Python

Libraries/Tools: Pandas, Numpy, Scikit-learn, NLTK, TfidfVectorizer, Matplotlib, Seaborn

Techniques: Sentiment Analysis, Logistic Regression, Support Vector Machine (SVM), F1-Score Evaluation, Cross-Validation

Outcome: Based on the analysis, the SVM model is recommended for production use due to its superior overall accuracy and handling of neutral sentiment

View Project

Predicting Customer Churn

Objective: Developed models to predict customer churn, enabling proactive retention strategies.

Technologies Used:Programming Languages: Python Libraries/Tools: Pandas, Matplotlib, Seaborn, Scikit-learn Techniques: Logistic Regression, Decision Trees, Exploratory Data Analysis (EDA), Data Cleaning

Outcome: Achieved 82% model accuracy and identified key churn predictors, improving retention by 25%.

View Project

Home Value Predictions

Objective: Predict home values in King County based on multiple features, such as size, location, and property age, to provide actionable insights into the real estate market.

Technologies Used:

Programming Languages: Python

Libraries/Tools: Pandas, Matplotlib, Seaborn, Scikit-learn

Techniques: Linear Regression, Exploratory Data Analysis (EDA), Data Cleaning

Outcome: Achieved high model accuracy, enabling practical use for real estate investors and stakeholders.

View Project

Microsoft Movie Studio

Objective:Analyze the Box Office Mojo and IMDB datasets to identify movie genres that perform well at the box office, providing actionable recommendations for investment and studio collaborations to increase box office success.

Technologies Used:

Programming Languages: Python

Libraries/Tools: Pandas, NumPy, Matplotlib, Seaborn, Scikit-learn

Techniques: Data Cleaning, Exploratory Data Analysis (EDA), Data Visualization, Statistical Analysis

Outcome: Provided actionable recommendations for investment in high-performing genres and strategic collaboration with leading studios to maximize box office success.

View Project

Titanic Survival Prediction

Objective:To understand the characteristics of those who survived to provide more insights for improving survival rates in analogous scenarios

Technologies Used:

Programming Languages: Python

Libraries/Tools: Pandas, NumPy, Matplotlib, Seaborn, Scikit-learn

Techniques: Data Cleaning, Exploratory Data Analysis (EDA), Data Visualization, Statistical Analysis

Outcome:I managed to predict Titanic survival correctly for 72% of people -as per this Kaggle competition results.

View Project

Kwanza Tukule: Sales Performance Analysis

Objective: To analyze sales data and customer purchasing behavior to provide business insights that improve decision-making.

Technologies Used:

Programming Languages: Python

Libraries/Tools: Pandas, Plotly, Dash, Matplotlib, Scikit-learn

Techniques: Data Cleaning, Exploratory Data Analysis (EDA), Data Visualization, Customer Segmentation

Outcome: The project identified best-selling products, peak sales periods, and segmented customers based on purchasing behavior. The insights support inventory planning and targeted marketing strategies.

View Project