The value of data-driven Customer Value Management or CVM cannot be underrated. Data and other algorithms/analytics that shape data are an imperative part of customer value management in a telecom company. With enhanced customer expectations, it is up to the ability of telecom companies to provide customers with a seamless experience and to also ensure that they help boost revenue in the process. To understand this concept in a more functional manner, I recently interviewed the chief of CVM at Mahindra Comviva, Amit Sanyal. With so much on hand to discuss, I got to the crux of the matter straightaway and asked Amit about the pillars he considered to be important for a customer value management program being driven by analytics. The prodigy responded to my questions by commenting that all methods of CVM being driven by the force of analytics should be dedicated towards these three pillars.

Where do we look to see the most advanced chatbots and the most complete application of AI? Chatbots designed as ‘artificially intelligent psychological counseling chatbots’, ‘therapeutic assistants’ for short.

We use if statements in our everyday life all the time – even if our everyday life is not written in Python. If the light is green then I’ll cross the road otherwise I’ll wait. If the sun comes up then I’ll get out of the bed otherwise I’ll go back to sleep. Okay, maybe it’s not this direct, but when we take actions based on conditions, our brain does what a computer would do: evaluate the conditions and act upon the results. Well, a computer script doesn’t have a subconscious mind, so for practicing data science we have to understand how an if statement works and how we can apply it in Python!

When dealing with forecasting models there is an issue that generates a lot of confusion, which is the difference between direct and recursive forecasts. I believe most people are more used to recursive forecasts because they are the first we learn when studying ARIMA models.

With the release of the new version of mlrMBO we added some minor fixes and added a practical feature called Human-in-the-loop MBO. It enables you to sequentially
• visualize the state of the surrogate model,
• obtain the suggested parameter configuration for the next iteration and
• update the surrogate model with arbitrary evaluations.
In the following we will demonstrate this feature on a simple example.

We are excited to announce the availability of the cloudml package, which provides an R interface to Google Cloud Machine Learning Engine. CloudML provides a number of services including:
• Scalable training of models built with the keras, tfestimators, and tensorflow R packages.
• On-demand access to training on GPUs, including the new Tesla P100 GPUs from NVIDIA.
• Hyperparameter tuning to optmize key attributes of model architectures in order to maximize predictive accuracy.
• Deployment of trained models to the Google global prediction platform that can support thousands of users and TBs of data.

Two shinydashboard posts in a row, that’s a first. As I mentioned on Twitter, I’m not really this productive; rather, the apps had been on the proverbial shelf for a while and I’m just releasing them now. In fact, this is one of my earlier works: quantifying the gender imbalance as it manifests itself in the LSE International Relations (IR) reading lists. You can access the app here. This is a much larger project that I got involved during its second year, so I’m just visualising other peoples’ hard work.1 The recentness of my contribution to the project was clearly on display when I amused my audience by saying cross-sectional feminism instead of inter-sectional. Are you a statistician or what? Baby steps. In a nutshell, about twenty or so PhD candidates at the department manually (!) scraped the reading lists of 43 courses that were on offer during the 2015-2016 academic year, resulting in a dataset containing 12,358 non-unique publications.2 Of those, 2,574 involves at least one female author, while 9,784 features at least one male author. Morevoer, 81% of the syllabi is written exclusively by male scholars.

Long short-term memory (LSTM) networks have been around for 20 years (Hochreiter and Schmidhuber, 1997), but have seen a tremendous growth in popularity and success over the last few years. LSTM networks are a specialized type of recurrent neural network (RNN)—a neural network architecture used for modeling sequential data and often applied to natural language processing (NLP) tasks. The advantage of LSTMs over traditional RNNs is that they retain information for long periods of time, allowing for important information learned early in the sequence to have a larger impact on model decisions made at the end of the sequence. In this tutorial, we will introduce the LSTM network architecture and build our own LSTM network to classify stock market sentiment from messages on StockTwits. We use TensorFlow because it offers compact, high-level commands and is very popular these days.