Robots have been a mainstay in factories for decades, but their use has been expanding everywhere else, from warehouses and hospitals to retail. That trend continued last year, and the novel coronavirus pandemic has accelerated service robot demand for automated logistics, disinfection, and delivery, according to the International Federation of Robotics. The Frankfurt, Germany-based IFR said that the sales value of professional service robots increased by 32% to $11.2 billion (U.S.) worldwide between 2018 and 2019. The organization published its full research in the "World Robotics 2020 – Service Robots" report, which is available for download. Sales of medical robotics accounted for 47% of the total service robot value turnover in 2019, according to the IFR.

Analyzing time series can be an extremely useful resource for virtually any business, therefore, it is extremely important for data scientists entering the field to have a solid foundation on the main concepts. Luckily, a time series can be decomposed into its different elements and the decomposition of the process allows us to understand each of the parts that play a role in the analysis as a logical component of the whole system. While it is true that R has already many powerful packages to analyze time series, in this article the goal is to perform a time series analysis --specifically for forecasting-- by building a function from scratch to analyze each of the different elements in the process. There are several methods to do forecasting, but in this article, we'll focus on the multiplicative time series approach, in which we have the following elements: Seasonality: It consists of variations that occur at regular intervals, for example, every quarter, on summer vacation, at the end of the year, etc. A clear example would be higher conversion rates on gym memberships in January or a spike in video game sales around the holidays.

When a data scientist/machine learning engineer develops a machine learning model using Scikit-Learn, TensorFlow, Keras, PyTorch etc, the ultimate goal is to make it available in production. Often times when working on a machine learning project, we focus a lot on Exploratory Data Analysis(EDA), Feature Engineering, tweaking with hyper-parameters etc. But we tend to forget our main goal, which is to extract real value from the model predictions. Deployment of machine learning models or putting models into production means making your models available to the end users or systems. However, there is complexity in the deployment of machine learning models.