Application fields of R in classical industrial analytics
0 likes321 views
The document discusses the distinctions between data science and classical analytics in industrial applications, particularly emphasizing the use of R programming. It highlights various applications of both fields, including predictive maintenance and statistical process control, while also addressing the capability of R to tackle typical industry challenges. The document calls for closer collaboration between R developers and quality management, further training of engineers, and the development of new R packages.
Application fields of R in classical industrial analytics
- 1. © 2010 – 2016 eoda GmbHAndreas Wygrabek Application fields for in classical industrial-analytics industry eRum 2016 Andreas Wygrabek Data Scientist
- 2. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Objectives 1. Differentiation of data science and classical analytics tasks in industry 2. Application scenarios for R in industry 3. Reveal the potential of R in scenarios of classical analytics What is this talk about?
- 3. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Representative applications of ... Data Science
- 4. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de AnalyticMetrological Data Data Science Data Science | Energy Forecast Weather Forecast
- 5. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de AnalyticSensor Data Data Science Data Science | Predictive Maintenance Machine Failure Predictions
- 6. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de AnalyticCustomer Data Data Science Data Science | Scoring Conversion Probability
- 7. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Representative applications of ... Classical_Analytics
- 8. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de AnalyticAcceptance Sampling Analytic Classical Analytics | Incoming Goods Inspection Results (Acceptance, Rejection)
- 9. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Analytic Sampling Data Analytic Classical Analytics | Statistical Process Control Process Charts
- 10. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Analytic Sampling Data Analytic Classical Analytics | Design of Experiments Cause-Effect-Relations
- 11. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de How do the applications of data science and classical analytics in industry differ?
- 12. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Exploratory Data Analysis (Data Science Methods) Predictive Classification Regression Clusteranalysis Modelling PCA Forecasts Data-Mining Algorithms Statistics Mu Sigma alpha/beta ProbabilitiesNormal Distribution Confidence Inference Statistics Descriptive Statistics Mean Range Variance Correlation … … Time Series … Testing Deep Learning Analytics and Data Science | Methodology Analytics Data Science
- 13. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Analytics and Data Science | Project Management Analytic: Traditional PM - Solution known - Standardized Issues - Ressource estimation Data Science: Agile - Interative solution - Individual Issues
- 14. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Analytics and Data Science | Data Analytic: Data is certain - Data certainly contains all required informations Data Science: Data is uncertain - Often it is not certain that the data contains the required information
- 15. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Analytics and Data Science | Difficulty Analytic is complicated - Experts will solve the issue Data Science is complex - No guarantee
- 16. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Analytics and Data Science | Location Analytic: Data Science: - Quality Management - Production - IT - Business Intelligence - Data Science Lab
- 17. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Analytics and Data Science | Implementation Analytic: Choosing the right tool Data Science: Proof of Concept
- 18. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de How capable is R to solve the issues which are typical in industry?
- 19. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de - Incoming Goods Inspection - Sampling - AcceptanceSampling - Statistical Process Control - spc - qcc - IQCC - qualityTools - SixSigma - DoE - CRAN Task View: Design of Experiments (DoE) & Analysis of Experimental Data - Most used package: agricolae Classical_Analytics Reporting - knitr - shiny - interfaces to external software in
- 20. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de Mind the gap 1. R is powerfull – all the tasks mentioned can be solved 2. R is free 3. R is able to report 4. The need for new tools in classical analytics is growing
- 21. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de What are the reasons? 1. Qualification 2. Old processes 3. Established Software
- 22. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de To Do´s 1. Bring R-Developers closer to QM and production 2. Upskill engineers in R 3. Further package development (AQL, ISO 2859, ISO 3951) 4. Create stable R environments
- 23. © 2010 – 2016 eoda GmbHAndreas Wygrabek www.eoda.de @eodaGmbH @eodaGmbH eodaGmbH blog.eoda.de eoda GmbH Universitätsplatz 12 34127 Kassel - Germany www.eoda.de/en info@eoda.de +49 561 202724-40 The Data Science Specialists.