It's copied to your library with every feature (audio, generate more, ask the teacher) without spending AI again on flashcards/quizzes. "Add the whole unit" also brings in the other topics in this category.
Data Science and Artificial Intelligence
University
Introduction to Data Science. Source: Wikipedia (CC BY-SA 4.0).
Sign up to open «Data Science and Artificial Intelligence»
Browsing the syllabus and its units is open to all. Opening the full topic — summary, quiz, flashcards and audio — is free but requires an account, so we can save your progress and keep this sustainable for everyone.
What you need to know
Read the full theory →Definition, nature, and profile of the data scientist
- Data science: an interdisciplinary field using mathematics, statistics, computing, and engineering methods
- Objective: to collect, process, analyze, and present noisy, structured, and unstructured data
- Nature: can be a science, paradigm, method, discipline, workflow, or profession
- Key component: integration of domain knowledge with computer science and mathematics techniques
- Fourth paradigm: data science is the data-driven paradigm (alongside the empirical, theoretical, and computational)
- Functions of the data scientist: collection, cleaning, exploration, modeling, visualization, and machine learning implementation
- Tools of the data scientist: programming code and statistical knowledge
- Professional background: mathematicians, engineers, economists, actuaries, physicists, chemists, and doctors
Historical evolution and definitions of Data Science
- 1962: John W. Tukey defines data analysis as procedures for analysis, interpretation, and planning.
- 1977: Tukey proposes the use of data to suggest hypotheses in statistical models.
- 1960s: Peter Naur uses 'Data Science' as a substitute for computational sciences.
- 1977: The IASC seeks to relate statistical methodology, computational technology, and expert knowledge.
- 1996: Formal use of the term in the IFCS conference in Kobe, Japan.
- 1997: C.F. Jeff Wu defines statistics as a trilogy: collection, analysis/modeling, and decision-making.
- 2001: William S. Cleveland establishes 6 technical areas: multidisciplinary research, models/methods, computing, pedagogy, tool evaluation, and theory.
- 2002: CODATA begins publishing the Data Science Journal.
- 2003: Columbia University publishes The Journal of Data Science.
- 2005: The National Science Board defines data scientists as experts in computing, software, databases, and disciplinary experts.
- 2008: Jeff Hammerbacher and DJ Patil adopt the term for their roles at Facebook and LinkedIn.
- 2009: Zhu and Xiong define Datalogy and Data Science as the study of data on the network.
- 2013: Creation of the IEEE Task Force on Data Science and Advanced Analytics.
- 2014: First IEEE international conference on Data Science and Advanced Analytics.
- 2015: Launch of the International Journal on Data Science and Analytics by Springer.
Applications of Data Science in Marketing, Governance, and the Data Scientist Profile
- Marketing: Use of databases to predict purchases and personalize messages.
- Spotify: Analyzes data from 170 million users for personalized recommendations.
- Netflix: Analyzes habits, connection times, and viewing order for 120 million users.
- Governance: The IDB promotes data science for public policies in Argentina and Brazil.
- Governance objectives: Mobility, smart cities, security, privacy, and accountability.
- Big Data: Massive data that exceeds the processing capacity of traditional applications.
- Data science: Analysis of disorganized data for decision-making.
- Data scientist: A hybrid profile of statistician, computer scientist, mathematician, and creative.
- Functions of the scientist: Value extraction, communication of results, and cost/benefit optimization.
Profile, skills, and challenges of Data Science
- Data processing: includes statistical inference, regression, and hypothesis testing
- Nathan Yau profile: requires statistics, APIs, databases, and extraction
- Fry's process: consists of 7 phases (acquire, analyze, filter, extract, represent, refine, and interact)
- Specialization risk: the fragmentation of approaches can cause information loss during transitions
- Davenport Profile: a hybrid of hacker, analyst, communicator, and advisor
- Key requirement: freedom to experiment and explore
- McKinsey Demand (2011): projection of 440,000 to 490,000 jobs for 2018
- Volume challenge: requires new methods for storage and integration
- Interoperability challenge: need for layers to connect different systems
- AI challenge: difficulty in the interpretability of models
- Cleaning challenge: tools to separate noise from valuable data
Flashcard preview (6 of 20)
- 01
Data science is an [____] field of knowledge
interdisciplinary
- 02
According to Jim Gray, data science is the [____] paradigm of science.
fourth
- 03
In addition to statistics and computer science, what other knowledge must data science integrate to analyze real-world phenomena?
application domain knowledge
- 04
What types of data (according to their structure) does data science process?
noisy, structured, and unstructured
- 05
Which professional uses programming code and statistics for data collection, cleaning, and modeling?
data scientist
- 06
In 1962, who preceded the term “Data Science” in their article “The Future of Data Analysis”?
John W. Tukey
+ 14 more flashcards, 12 quiz questions and audio when you sign up (free).