Introduction to data science and IoT
1 like193 views
This document provides an introduction to data science and IoT. It discusses what data science is, the main academic disciplines that constitute it (including mathematics, artificial intelligence, statistics, and machine learning). It also defines data and the different types and structures of data. The document introduces big data and the 5 V's of big data. It discusses the data science process and some common applications of data science like recommender systems and predictive maintenance. Finally, it provides an overview of IoT platforms and how predictive modeling and analysis can be used for maintenance in IoT applications.
Introduction to data science and IoT
- 1. El khadir LAMRANI Introduction to Data Science & IoT
- 2. Plan 2 • Introduction to data science • What is data science ? • Data ? • Big data ? • Algorithms • Data science process • Data science application • Data science for IoT • IoT Platform • Predictive maintenance modeling & Analysis
- 3. Introduction to data science What is Data science ? 3 • An interdisciplinary field about scientific methods, processes, and systems to extract knowledge or insights from data in various forms • Concept to unify linear algebra, statistics, dataanalysis and their related methods in order to understandand analyze actual phenomena with data • An empirical approach that relies data to provide an answer to problems. Mathematics Artificial Intelligence Domain expertise Computer vision Statistics Machinelearning Advancedcomputing Linguistics Database& StorageVisualization Main academic disciplines that constitute data science
- 4. Introduction to data science As the beginning was data .. juin 18 4 What is data ? “Data is the New Oil” – World Economic Forum 2011
- 5. Introduction to data science As the beginning was data .. juin 18 5 • Type of Data • Quantitative • Qualitative • Complexity of data • Structured(relational database) • Semi-structured(Json, XML, CSV, Logs) • Unstructured(Images,Texts, Videos, Signal, ..) • Where are the data ? • Everywhere • Open dataset • API (twitter,google, Wikipedia,..) • Web (web scrapping) • IoT (sensors) • …
- 6. Introduction to data science Big data juin 18 6 5 Vs of Big Data • Raw Data: Volume • Data types: Variety • Change over time: Velocity • Data Quality: Veracity • Information for Decision Making: Value
- 7. Introduction to data science Big data juin 18 7 Hadoop Ecosystem -Hortonworks data platform
- 8. Introduction to data science Algorithms.. To do what ? 8 1. Discover links in the data 2. That’s all
- 9. Type of problem to resolve • Classification • Regression Learning mode • Suppervised Regression/Classification • Unsuppervised Data driven (Clustering) • Reinforcement Algorithms learn to react to an environment Introduction to data science There is a taxonomy of algorithms 9
- 10. Famous machine learning algorithms (Source: SAS Blog, Hui Li) Introduction to data science There is a taxonomy of algorithms 10
- 11. Introduction to data science Data Science Process 11
- 12. • Transaction Databases -> Recommender systems (NetFlix), Fraud Detection • Wireless Sensor Data ->Smart Home, Real-time Monitoring, Internet of Things • Text Data, Social Media Data -> Product Review and Consumer Satisfaction (Facebook, Twitter, LinkedIn), Chatbots, .. • Genotype and Phenotype Data -> Patient-CenteredCare, Personalized Medicine, .. • Images, vidéos -> Health-care, Facial recognition, Introduction to data science Data science Applications 12
- 13. Vast network of devices connected to the Internet • Remote Monitoring of things. • Things collect and exchange data. Data science for IoT What is IoT ? 13
- 14. Introduction to data science IoT Platform 14
- 15. Data science for IoT 15 IoT Platform Capteurs Passerelle IoT IoT HuB Azure Stream analytics Event HuB Données de référence Base de données Tableauxde bord Fichier JSON Fichier JSON Modèle de prédiction Fichier JSON Prédiction en temps réel Fichier CSV Kafka KafkaSpark streaming
- 16. Techniques are designed to help determine the condition of in- service equipment in order to • Build a model that predicts yield failure on their manufacturing process • Through analysis determine the factors that lead to yield failures in their process. Data science for IoT Maintenance predictive modelling & analysis 16
- 17. • Our system will be able to • Predict failures • Classify failures • Find similar cases • Analyzing the reasons of failures • Recommend a solution Data science for IoT Maintenance predictive modelling & analysis 17
- 18. Thank you. 18