Finally, I reached the final project for this self-paced data scientist program. As a final capstone project, I chose to develop a recommendation system on Netflix shows, which I have not conducted before.
My major in university is Biomedical Engineering, in which medical image processing is a major field. I have previously learned image segmentation, image reconstruction, and image classification, etc. However, I have not used deep learning models in these analyses before, since deep learning model has not been proposed when I am a student in the collage. And I am glad to have a chance to do medical image classification using deep learning model.
This blog covers the steps to analyze SyriaTel customer data. In this project, I tried to find the features that are significantly related to the churn of customers for SyriaTel service. To this end, I have both analyzed individual features with respect of the churn and built binary classification models for prediction of customer’s churn label.
As a data scientist, it is very important to choose a proper method to viusalize the data. The use of different visualizaiton methods have different purposes and could reveal the data in different aspects. The good visualization could make the communication with the stakeholders more effectively. Here, I introduced two commonly used visualization methods in data science.
I have been working with Data for over 10 years, and the data I mentioned is the brain signals that are recorded from many human participants either healthy or patients with some kinds of diseases, e.g., autism. And from these data, I am developing analytic methods to dig patterns or biomarkers that can be used to understand how our human brain is working and/or why certian brain diseases occured. However, I didn’t have comprehensive training in data science, but learned skills by empirical learning via internet or literature. Therefore, my skill is still quite limited.