Jose Dominguez - A Homespun Decentralised DIY Data Science Research Pipeline for IoT скачать в хорошем качестве

Jose Dominguez - A Homespun Decentralised DIY Data Science Research Pipeline for IoT

"A Homespun Decentralised DIY Data Science Research Pipeline for the Internet of Your Things" So many buzzwords; so little time. What do data science, digital signal processing, machine learning, databases, mobile platforms, devops, containers, sensors, and micro-controllers have in common? You’ve heard it a million times, it is so easy to collect your own data with things such as Arduino sensor shields, wristbands, and even mobile phones. You can monitor your sleep, the time you spend in the pub, count how many steps you took last Sunday, and how many times you listened to Taylor Swift last week, even though your twitter bio reads that your favourite band is Radiohead. But when you use commercial products, what happens to your data? where does it all go? and more importantly, how is it used? There is no final answer to those questions but some of that data can be used to derive a lot of your personal information, as Liccardi et al. demonstrate in I know where you live: Inferring details of people’s lives by visualizing publicly shared location data, ACM SIGCHI 2016. This talk is about how to construct your own data pipeline, from collection, to signal processing, to storage, and to aggregated analysis and visualisation. Through a series of examples, a number of basic data science principles will be explained, showing what can be achieved with data, and certain concerns that it brings in terms of privacy. A number of free and open source tools will be used throughout the examples, with a focus on tools that adhere to End User Development (EUD) principles, from data gathering with a drag and drop, blocks based, mobile app development tool, to routing data through a flow based system, and to displaying charts and graphs in an interactive notebook.