Velib bicycles status using api

Velib Bike Status using API is my Smart City Project. The aim of this project is to explore Velib public renting bicycle sharing system in Paris by using the API concept to display real time statues of the bicycle in the entire Paris region. The project uses pandas, urllib.request, folium, and RESTful APL to retrieve real time bicycle location, number of bicycles available etc.

• Date: November 2022

Injury prediction for athletes

The aim of this project was to develop a machine learning model that can predict injuries based on the training regimen of an athlete. It can be used to make injury prevention approaches easier to develop. This project focuses on Machine Learning Classification analyses, and it uses two (2) separate datasets, namely, the weekly & daily. The dataset was downloaded via Kaggle platform.

The project prediction was based on three (3) models on the weekly approach dataset namely, the DNN, K-NN, and XGBoost algorithm while on the daily approach, the project also implemented four models, namely, DNN, K-NN and XGBoost. Thus, the best-performing model on the daily dataset was K-NN with an accuracy of 97%. The weekly dataset on the other hand, k-NN was the best model with about 97% accuracy (f1-score). Furthermore, the feature with the highest influence on injury status was the ‘sum of max km’ with about 4.7 to 34.5km per day according to explorations on the data.

• Date: October 2022

Forecasting foreign exchange rate

This is my 'Hamoye AI Lab' Data Science summer internship collaborated premiere project. This project focuses on Time Series analyses. The dataset used for this project was downloaded on Kaggle. The dataset contains some currencies with their rates in dollar units.

The project used three models (Linear Regression, ARIMA and FBprophet Algorithms) and these models’ performances were evaluated using different time series forecasting performance measures by testing how well these models have predicted with respect to the actual observations. The project also compared and determined which one of these models will be able to make a better forecast comparing the results from the performance evaluation methods. In conlusion, the 3 models we built, ARIMA Model gave us the best results with the lowest Evaluation Matric of MAE,MSE,RMSE and MAPE respectively. This was used to forecast the Exchange rate for the last 30 days.

• Date: August 2022

Development of smart healthcare solution

Over the years, life expectancy has increased with accelerated aging of the population, which has called for increased attention. Elder care and medical issues are becoming a major social issue in the world.

Thus, this project implemented a full-fledged IoT connected system that aimed as collecting information from the sensors layer and send it through gateway (raspberry pi). The platform allow the hospital users to check their vital signs (e.g pulse, temperature) and other ambient values such as temperature and humidity. Some functionalities may raise alerts (sensors threshold exceeded for instance) at the visualisation level.

• Date: June 2022

Pong game using flex Sensor paddles and twitter api integration

This is my Electronic of Object with regards to Internet of Things (IoT) project. The project focuses mainly on the design on both software and hardware interaction framework of Pong Game using Flex-sensor Paddles. The project Integrated Server/Client Interaction and Twitter API Communication.

The latency between Client/Service communication were thoroughly measured using three different mobile network services, such as 3G, 4G and 5G.

• Date: May 2022