Eric Schmidt, the former Google CEO revealed that every two days we generate as much information as we did from the dawn of civilization up until 2003 - that’s an estimated 5 exabyte of data every 48 hours. Most of the new data sources according to Eric Schmidt emanates from user-generated content is in the form of instant messages, tweets, photos et cetera. As a result, global companies and organizations have found themselves wondering how to use internal and external data to generate actionable insights and discoveries that can drive growth. It is with this in mind that iHub Data Lab hosts the yearly data science jams to train practitioners in the data science field.
Since the training's conceptualization in August 2013, the event has seen more participants take part in the field of data science. This year, iHub held it's fourth annual data science jam. The event kicked off with a two-day event by Software Carpentry with the theme 'Computational Methods for Unstructured Data.'Jason K. Moore from software carpentry took the attendees through a series of introductory lessons covering the unix shell, Version Control with Git, Managing Data with SQL and R programming. Those interested in taking the course further enrolled in a five-week training that took place at the UX lab every evening. They went through a series of modules ranging from introduction to machine learning to visualization and the use of algorithms to find patterns in large datasets.
The lessons were conducted at the iHub UX lab every evening from 4 to 7 pm owing to the fact majority of the attendees worked during the day. Some of the companies they represented include KopoKopo, Sanergy, KEMRI, and ipsos. Each day they went through a specific machine learning module. All of which covered four main techniques; supervised, unsupervised, reinforced and ensemble learning. Machine learning is based on the idea that instead of writing programs that solve the problems, we write the programs that learn to solve the problem.
For example, supervised learning is synonymous to guiding a child on his day-to-day activities. Unsupervised learning would be leaving a child on his own - they would pick a needle and pierce themselves,get injured and from this seek an alternate toy to play with. Therefore, their knowledge on surviving the world would come from a set of experiences. Reinforced learning would be similar to giving your dog a biscuit every time it pees at the right place triggering a reward mechanism. Finally, ensemble learning borrows from all the mentioned types of machine learning techniques. All these methods have been at the core of data science and it's what differentiates it from other types of programming. In data science, you write the program to learn solution, you don't hard code the rules.
iHub Data Science Lab offers consultancy services to startups and corporates looking to expand into new markets. This is done in conjunction with the UX lab, Research and Software and Design teams. As we mark the end of the 2016 data science jam, we invite you to the graduation ceremony of this year's cohort of Data Science students on Wednesday, August 17th, 2016 at the iHub (4th) from 5:30 pm.
For any queries reach us on [email protected]