Big Data Analytics

In the previous post, we looked at some basic models that are used for channel attribution. As discussed earlier, more refined models are required for an objective budget allocation. Here, we’ll look at two recent methods that take into account the complete conversion chains. We’d also discuss about A/B lift,... [Read More]

With the advent of data driven and measurement-focused marketing, cross channel attribution has become a hot topic. Briefly, it is used to assign proportional credit to each marketing touch point be it offline or digital that leads to a desired customer action. Therefore, attribution analytics can help the brands to... [Read More]

Helpful Resources for Workshops [Read More]

As a data scientist, you’d need to learn several tools and technology stacks. The learning curve required for each of those can get overwhelming. On top of it, each tool and technology would require some familiarity with its own default editor. It is difficult to concentrate on a singular task... [Read More]

In the previous post, we looked into generating synthetic EMG signals. In this post, we’ll look into feature engineering techniques. In particular, we’ll run into feature extraction techniques that will be useful in separating anomalies from normal signals. We’ll start this section with a discussion about time-frequency domain analysis. Since,... [Read More]

A few years back, I was working with some people at Yale University, Orthopaedics and Rehabilitation Department. We were working on a data set of electromyographic signals collected from three hundred and three varsity athletes recruited for a study on low pack pain. In this series of blog post, I’ll... [Read More]

Reading some of the content shared, I assume that the study is intended to provide interpretable results and variable importance in predicting Zdyad_msi_globdist. This follows that instead of looking for the best model, I’d initially look into finding the most interpretable interactions using the provided features. Starting off, let’s look... [Read More]

The primary purpose of this study is to identify commercial areas from residential areas using the density of businesses present in a certain area. For this, the spatial information in the Yelp data set was analyzed to garner areas of interests that identify some hypercontext awareness of the locality. At... [Read More]

In an exploration of some data on open street maps (OSM), FourSquare, and analysis of the same places on ground showed some discrepancies between addresses, names and actual location. In other words, the same place had different names at different data sources and hence a need for normalization is required.... [Read More]

A GTFS feed is a collection of csv files that stores a transit system’s scheduled operations as visible to riders. The specification is designed to be sufficient to provide trip planning functionality. An excellent visual explanation by OpenPlans of this system is provided in the figure below: [Read More]