Blogs

With the coming of age of Artificial Intelligence, everyone wants to see how they can use their data to solve their business problems. For example, some Brick and Mortar stores are silently working on how they can stay competitive by predicting empty supermarket shelves. This leads to the idea that the store can automatically track which items are being pulled off of the shelves, and take the guesswork out stocking. With the advancements and neural networks and artificial intelligence, I wanted to see if we could predict the empty spaces on supermarket shelves, given a limited amount of data and no initial bounding boxes to train on.

Everyone knows high-level languages are wonderful. The issues of yester-year are long gone; we no longer have to build dashboards from scratch. Being able to query data, and display directly through R is here, and I don’t see it going anywhere anytime soon. Flexdashboard is an open source package in R that allows users to display dashboards directly in browser. It also has publishing support through RStudio. Assuming you have moderate traffic (like me), it’s perfect!

In the United States, various factors affect the airfare prices that consumers see. An issue is that a single linear model may not fit the data, as prices can vary drastically across different routes. We will obtain Quarterly US Air Fare and Volume data and utilize a mixed effects model to predict air fare prices. The original dataset contains 108,602 observations and 14 attributes and there are no missing values. For the sake of simplicity, in this paper, we only consider a random sample of 50 routes as the original dataset contains 4177 routes.

Community colleges for many is a stepping stone to a larger dream of four year university. However, many of the risk factors that prevent students from succeeding in school are highly prevelant in the student body of 2-year institutions. In this post, we explore open college data. This dataset includes the name of the college, the aggregated number of people who passed a particular subject by transfer level and race.

Although there are many examples, churn prediction is one of the classical applications of Data Science that works. Churn prediction gives businessmen and bussinesswomen the power to catch those consumers who are likely to leave the company. They can in turn take appropriate measure to keep their business. In this project, we consider a dataset of 7,043 observations. The goal here is to predict, with a fair amount of accuracy, which observations are likely to churn.

R Markdown supports Python integration with the package reticulate.The nice thing about R Markdown is that it is the first scripting language I have ever learned. Therefore, I am almost too comfortable with it. It is my go to for most of my project because it’s so easy to get a proof of concept with new ideas. The downfall of it - until now - is that I was unaware that we could use PYTHON IN R MARKDOWN!