Projects
Overview: The Tanzanian Ministry of Water collected data across Tanzania that identified which pumps are functional, need some repairs, and don't work at all. Utilizing classification techniques, I was able to understand which water points are functional or need repairs. These findings can improve maintenance operations and ensure clean; potable water is available throughout Tanzanian communities.
​
The technique of solving: Multiclass Classification
​
Overview: Emotions play a crucial role in human interaction and understanding them can enhance various applications from customer service to mental health. Utilizing the GOEmotions dataset, I developed a classification model using BERT to accurately detect 27 distinct emotions. To make this model interactive and accessible, I integrated it with Streamlit, allowing users to input any phrase and receive the detected emotion in real-time. This project demonstrates the powerful combination of state-of-the-art natural language processing and user-friendly interfaces.​
The technique of solving: NLP Classification with BERT and Interactive Application with Streamlit
Overview: Planning a vacation can be daunting with the multitude of choices available. The Vacation Planner is an interactive application designed to simplify this process by providing personalized recommendations. Users can find the best times to visit their chosen destination, discover top attractions, locate ideal places to stay, identify local foods to try, and learn about events and shopping areas. Utilizing advanced natural language processing techniques, the Vacation Planner ensures that users receive customized suggestions based on their preferences. The app is powered by OpenAI's GPT model and seamlessly integrated with Streamlit for an intuitive and engaging user experience.​
The technique of solving: Sequence of NLP Chains with LangChain, OpenAI GPT Model, and Interactive Interface with Streamlit
Overview: Chatbots are widely used on websites like Amazon and eBay to provide instant customer support. In this project, I developed a customer service chatbot using Google Dialogflow to handle common customer inquiries. The chatbot can understand and respond to user questions about order statuses and details, utilizing natural language processing to interpret queries and extract necessary parameters to deliver accurate and helpful responses.​
The technique of solving: NLP with Dialogflow
Overview: Two consecutive quarters of negative GDP, war, and increased interest rates because of quantitative easing led us to one of the most interesting economic periods in history. I've done a time series for all urban consumers on all items except food & energy to dive deeper into the past, present, and future.​
The technique of solving: Time Series
Overview: A world without insects means no mosquito bites, but who would aerate the soil, help reproduce flowers, disburse nutrients, and plenty more ecological services? Well, the world might soon be without the monarch butterfly for various reasons, from habitat destruction to climate change. I did a multi-classification project on butterflies not because I'm a tree hugger or environmentalist but because I understand the impact of a world with no insects.​
The technique of solving: Multiclass Classification
Overview: From the Rastafarians to the food and the herbs Jamaica is definitely one of the most loved countries on earth. To me, the richest part of their culture is the way they make broken English sound good again. Take a look in my notebook and look at a text generator for the greatest dialect in the world.​
The technique of solving: NLP Text Generation
Overview: The management at a big supermarket chain collected data from their stores all over the country for a specific nondisclosed item. With the information given from the data, I predicted the sales per product for each store. With this information, each store can understand which stores perform well at selling the item and which stores do not.
The technique of solving: Regression
Overview: Bitcoin and other cryptocurrencies are changing how technologists, financiers, and economists view the world. The second biggest market capitalization cryptocurrency after Bitcoin is Ethereum. Ethereum is the largest open-ended decentralized software platform. Ethereum brings tremendous value to the future of technology and those that invest financially in its future. With this information, I predicted the closing price point of Ethereum over time.
The technique of solving: Time Series
Overview: Fashion-MNIST is a dataset from Zalando Research's article, which consists of fashion images. I used digital image processing to determine what clothing items are what. With this algorithm, clothing items represented by pictures in Zalando's article can be differentiated.
The technique of solving: Multiclass Classification
