Applied data science and engineering for local weather forecasts by Nikhil Podduturi скачать в хорошем качестве

Applied data science and engineering for local weather forecasts by Nikhil Podduturi

This is a record of Nikhil Podduturi's talk at the GOTO Berlin 2016. Abstract Progress in weather forecasting and in climate modelling over the past 50 years has been dramatic. Due to these dramatic improvements the data being generated increased exponentially. While the first computer ENIAC (Electronic Numerical Integrator and Computer) took 24 hours to make a 24 hour integration weather forecast, right now at MeteoGroup we generate weather forecasts for next 10 days on the fly from the browser. At the core of addressing this grand challenge is MeteoGroup’s engineering team. For assimilating and crunching huge amount of data, we build our own PaaS and enjoy leveraging the benefits of the microservice architecture and agile methodology. Description: In this presentation we would like to tell the story below: • How we use latest big data technologies and cutting edge engineering principles to crunch around 1 terabyte of data per day. • How we combine different fields of study i.e., technology, data science and meteorology to do the weather prediction. • Give you insights and share our lessons learned in e.g. testing varying • Data streams –––– Target Audience: • Data engineers • Data scientists • Weather enthusiasts –––– Links: • GOTO Talk: https://gotocon.com/berlin-2016/prese... • Slides: https://gotocon.com/dl/goto-berlin-20... –––– Nikhil Podduturi: Github.com: https://github.com/nikhilRP Github.io: http://nikhilrp.github.io/ Twitter: @nik_r_p