Here , I have implemented Logistic Regression using gradient descent. I have not used scikit-learn to do it. I did it so I could understand the inner intricacies of Logistic Regression which tends to be hidden when I use any library. I have mainly used numpy library for this task. I have used my own dataset on which I have performed preprocessing (again without using scikit-learn). I have compared my accuracy with scikit-learn at the end. And to my surprise I have beaten scikit-learn's accuracy.