Renée Michèle Niangoran

Brain Tumor Detection with AI

Data cleaning

Data preprocesing

Using GCP to store MRI images to train the models

Classification 2D with CNN in order to recognize the images with tumor (using the metrics recall)

Segmentation with Unet 2D in order to localize the tumor precisely (using the metric dice score)

Calcul the aera of the % of the brain tissue area

Insert the api (prediction model) in a web site that will be visible by the radiologue

Data analysis mission to improve sales without reducing customer satisfaction:

Fetching data from Mysql database (sqlite 3)

EDA

Data cleaning

Data processing

Data visualization ( Metabase )

statistical inference

Power point présentation to the CEO of the group on a solution to reduce customer dissatisfaction

- Using advanced computerized models to extract the data needed ( Python -SQL - Scrapping-API) - Cleaning dataset by removing corrupted data - Exploratory data analysis : 1. Initial analysis to assess the quality of the data 2. Performing further analysis to determine the meaning of the data 3. Performing final analysis to provide additional data screening 4. Statistical inferences 5. Communication - Feature engineering - Applying machine learning algorithms and statistical techniques to resolve business issues. - Applying deep learning algorithms - Developping pipelines and tools to monitor and analyse model performance and data accuracy - API creation (Fast API) - Integrate data from 3rd party services via ETL tools and custom pipelines (GCP) - Deployment to production ( Heroku - Streamlit -GCP ) - CI / CD