We have all used MS Paint at some point in our lives, but its capabilites are limited...or so you thought. Using PyAutoGui library one can automate any application. For now I've tried to automate MS Paint to draw concentric circles and sqaures. The possibilites are endless. The code and some example images can be found below:
Here
Visualizing K-Means Clustering
K-means Clustering algorithm is one of the fundamental algorithms used in machine (Unsupervised) learning. It enables researchers to visualize hidden clusters in a random set of data, which in turn allows them to convert that random data into powerful information. I have done the same but without the help of any Data Science Library. Use Cases are ever expanding. The code and some example images can be found below: Here
Perceptron in Action
Every Data Science enthusiast must have heard about Neural Networks and its capabilities. One needs to know that any function can be approximated to arbitrary accuracy by using two layers of such Neural Net. Perceptron on the other hand is a Linear Classifier with limited capabilities. Nevertheless, Here I have visualized Perceptron training without using any Data Science library. The code and some example GIF can be found below: Here
Deciding Production Lot Sizes
First thing to know about Capacitated Lot Sizing Problem is that it belongs to the class of NP-Hard problems. With that in mind the main objective is to determine the production lot sizes of every item on each machine per period, while satisfying the demand and minimizing the cost associated with it. I have implemented a small instance of this problem using DOCPLEX, a python library from CPLEX. The code, method used and some example solutions can be found below: Here
Automating Email Deletion
An Email ID serves as a medium of communication between you and various platforms/services you have signed up for, because of which your inbox gets filled with irrelevant promotions and newsletter. To tackle this problem I made a program using a python library Selenium to automate the process of deleting emails. The user will have to insert the sender's name in the master_list.json and the algorithm will take care of rest. The code can be found below: Here
Time Series Forecasting:
ARIMA Models
Time Series Forecasting is crucial for many companies and if done correctly it can be very rewarding. Based on Box-Jenkins Methodology, ARIMA models are very powerful Data-Driven models which can perform exceptionally well and yeilds reliable forecasts, but finding a suitable fit requires time and skill as it is an iterative process. It is said and I quote "Box-Jenkins ARIMA modeling is more an art than science." I have done my analysis considering such ARIMA models on RStudio as well as Python. Analysis Reports and Source code can be found below: RStudioPython
Hangman on Python
We have all played Hangman or as known in India 'Hollywood/Bollywood', at least once in our childhood. Thus, I decided to bring back the nostalgia by making the same game we used play in free periods. Thanks to OMDb's API serive, the best part about this built is the non-requirement of any movie database which helped to deploy the project online on a static page. Therefore with an active internet connection anyone can play this game. Click the Play button to be redirected to the game. (Note: Play on a bigger screen for better experience.) Play
Vizzing with Tableau
Tableau is a powerful data visualizing tool which enables users to create interactive dashboards or so called Viz in Tableau's environment. It is a must for any data professional for delivering effective reports. I have compared and summarized the growth of Video Games against Cinemas by taking 20 titles which generated highest revenue from the last 20 years. The result and the dashboard can be seen by clicking the link down below. Show Viz
