FIT1043-Week1.2-DataScienceProcess.pdf

FIT1043 Introduction to Data Science Week 1, Data Science Process Dr. Sicily Ting Fung Fung School of Information Technology Monash University Malaysia Data Science Process ePub Section 1.3 Learning Outcomes Week 1 By the end of this week you should be able to: Explain what is data scienc...

FIT1043 Introduction to Data Science Week 1, Data Science Process Dr. Sicily Ting Fung Fung School of Information Technology Monash University Malaysia Data Science Process ePub Section 1.3 Learning Outcomes Week 1 By the end of this week you should be able to: Explain what is data science and Drew Conway’s Venn diagram Comprehend the usefulness of machine learning Explain different components of a data science process Differentiate data science from other related disciplines Learn how to install and start coding in Python with Jupyter Notebook To be achieved in your tutorial / laboratory session The Data Science Process ePub Section 1.3 What happens in a Data Science project? Illustrating the process A quick walkthrough illustrating the steps The standard value chain Our model of the process The Data Science Process Illustrating the Process Many different tasks come together to complete a Data Science project A data scientist should be familiar with most, but doesn’t need to be an expert in all Not all are labelled as Data Science Some from other field such as computer engineering, business,... image source: “Young Business Man Holding a Tablet” by Pic Basement 1. Pitching ideas: For data science projects to investors / managers image src: Stephen Ausmus acquired from USDA ARS, public domain. 2. Collecting data: Researchers preparing to x-ray a patient.  3. Integration: Data can come from many different sources. icons from by Openclipart.org, public domain  4. Interpretation: Data can be described using a database schema. image source: Eric, Sql Designer  5. Governance: (i) caring for the data and its subjects. (ii) managing data standards and formats. icons from by Openclipart.org, public domain; Good and Evil by AJC ajcann.wordpress.com  6. Engineering: Data engineers make the back-end work image source: by Intel Free Press  7. Wrangling: Inspecting and cleaning the data. image src: “rstudio” by mararie  8. Modelling: Proposing a conceptual / mathematical / functional model. image source: “Mathematics” by Tom Brown 8. Modelling: Analyst building models with his favourite tool.  8. Modelling: Analysis, statistics and/or machine learning works on the data. image source: “From Data to Wisdom” by Nick Webb  9. Visualisation: Visualising data to interpret it and present results. image source: Stephen Ausmus acquired from USDA ARS, public domain  9. Visualisation: Choosing appropriate visualizations for the data. Many different options exist! image source: “Visualization Matrix” cropped, by Lauren Manning  10. Operationalize: Putting the results to work. image source "Illustration of Strategy“ by Denis Fadeev  Data Science Process Standard Value Chain: Our Mode of the Process Slide 19 / Parts of a Data Science Project Collection Getting the data Engineering Storage and computational resources across full lifecycle Governance Overall management of data across full lifecycle We call this the Wrangling Standard Value Data pre-processing, cleaning Analysis Chain. Discovery (learning, visualisation, etc.) Visualization Arguing the case that the results are significant and useful Operationalize Putting the results to work, so as to gain benefits or value Data Science Process from Doing Data Science by Schutt and O’Neil, 2013, (available digitally through library) Chapter 1 of the book provides the following visualisation of the standard value chain for a data science project. A typical data scientist has a different mix of skills as well as domain knowledge Data Science Process from Doing Data Science by Schutt and O’Neil, 2013, (available digitally through library) Data Scientist Addresses the data science process to extract meaning / value from data Data Science Process from Doing Data Science by Schutt and O’Neil, 2013, (available digitally through library) Chief Data Scientist A form of chief scientist who addresses data management, data engineering and data science goals. Chief Scientist corporate position, responsible for science related aspects of a company/organisation  Relationship of Data Science to Other Disciplines http://growtrue.net/coursera- data-science-courses/ Slide 24 / Related: Data Engineering Building scalable systems for storage, processing data Hadoop, Databases, Distributed processing, Datalakes, Cloud computing, GPUs, Data wrangling, … Huge, continuous improvement.... Related: Data Analyst Performing analysis and understanding results R and Microsoft Azure Machine Learning Machine learning, Computational statistics, Visualisation,... Huge, continuous improvement.... Related: Data Management Managing data through its lifecycle ANDS Ethics, Privacy, Providence, Curation, Backup, Governance,... Huge, continuous improvement.... Home Activities Suggested Activities for the week end Videos Watch Cukier’s TED talk on “Big Data” Watch the CERN video, “Big Data” from Tim Smith Links to resources providing historical background to data science: Wolfram Alpha: computable knowledge history Cloud Infographic: Evolution Of Big Data The Web Technology timeline A brief history of Data Science What is Data Science? Recap: Learning Outcomes Week 1 By the end of this week you should be able to: Explain what is data science and Drew Conway’s Venn diagram Comprehend the usefulness of machine learning Explain different components of a data science process Differentiate data science from other related disciplines Learn how to install and start coding in Python with Jupyter Notebook To be achieved in your tutorial / laboratory session