obsidian/WS2425/Data Science/VL/Zusammenfassung.md

1

1. Organizational Information

   • Page: 2
   • Notes:
     • Contact: klaus.kaiser@fh-dortmund.de
     • Room: B.2.04
     • Professor Klaus Kaiser has a background in data science across various industries.

2. Introduction to Data Science

   • Page: 4-7
   • Notes:
     • Definition: Data science is about turning raw data into meaningful insights.
     • Interdisciplinary field combining statistics, computing, and domain knowledge.
     • Historical context of the term “data science” from 1962 to 2001.

3. What is Data Science?

   • Page: 8-10
   • Notes:
     • Data science involves using methods and systems to extract knowledge from data.
     • The intersection of math/statistics, computer science, and domain knowledge is crucial.

4. Practical Example of Data Science Project: Monkey Detection

   • Page: 16-24
   • Notes:
     • Steps include understanding the problem, data collection, labeling, model training, and deployment.

5. Related Fields in Data Science

   • Page: 14-15
   • Notes:
     • Data Engineering: Building systems for data collection and processing.
     • Data Analysis: Inspecting and transforming data to inform decisions.

6. Tasks in Data Science

   • Page: 16
   • Notes:
     • Overview of different tasks within classical machine learning.

7. Real-World Examples of Data Science Applications

   • Page: 25-32
   • Notes:
     • Applications include autonomous driving, face recognition, predictive maintenance, fraud detection, recommendation systems, and cancer detection.

8. Overview of Lecture Content

   • Page: 34-35
   • Notes:
     • Basic topics include data basics, statistics, presentation techniques, and machine learning.

9. Organizational: Schedule and Exam Information

   • Page: 38-40
   • Notes:
     • Lecture and exercise schedules, language of instruction, and exam details (written exam with bonus points for data analytics).

10. Expectations from Students

    • Page: 42-43
    • Notes:
      • Emphasis on respect, professionalism, and willingness to participate.

11. How to Continue in Data Science

    • Page: 46-48
    • Notes:
      • Suggested literature for further reading and related courses available in the curriculum.

12. Summary & References

    • Page: 51-55
    • Notes:
      • Key takeaways: ability to explain data science and recognize its applications.
      • Important references for further study are provided.

2

• Data Science Definition: Creating knowledge from data using math, statistics, and computer science.

• Data Types:

  • Structured: Follows a predefined model (e.g., tables).

  • Unstructured: Lacks explicit structure (e.g., text, images).

• Data Categories:

  • Discrete vs. Continuous

  • Nominal, Ordinal, Interval, Ratio

  • Qualitative vs. Quantitative

• Data Interchange Formats: Common formats include CSV and JSON.

• Data Trust: Importance of data quality dimensions: accuracy, completeness, consistency, timeliness, uniqueness, validity.

3

• Data Categories: Discrete, continuous, nominal, ordinal, interval, ratio, qualitative, and quantitative.

• Data Interchange Formats: Common formats include CSV and JSON.

• Data Quality Dimensions: Accuracy, completeness, consistency, timelessness, uniqueness, validity.

• Data Types: Primary (real-time, specific) vs. secondary (past, economical).

• Data Acquisition Methods: Capturing (sensors, surveys), retrieving (databases, APIs), collecting (web scraping).

• FAIR and Open Data: Principles for sustainable data usage and importance in scientific reproducibility.

4

• Primary vs. Secondary Data: Primary data is collected for a specific purpose, while secondary data is sourced from existing datasets.

• Data Collection Techniques: Includes scraping, which extracts data from websites, and considerations for legality and data protection.

• Data Protection: Emphasizes GDPR compliance, anonymization, and pseudonymization of personal data.

• Statistics Basics: Introduces descriptive and inductive statistics, frequency distributions, and graphical representations like histograms and bar charts.

• FAIR Principles: Focus on data findability, accessibility, interoperability, and reusability.

5

• Data Scraping: Extracts data from program outputs; should be a last resort.

• Anonymization: Removes personal info to protect identity; pseudonymization allows identification with additional info.

• Statistics Types: Descriptive, explorative, and inductive statistics.

• Frequencies: Absolute and relative frequencies; visualized through histograms, pie charts, and bar charts.

• Central Tendencies: Mode, median, and mean; box plots visualize data distribution.

• Statistical Dispersion: Measures spread of data; includes range, quartile range, and empirical variance.

6

• Histograms: Visual representation of frequency for continuous data.

• Cumulative Frequency: Measures total frequency up to a certain value.

• Statistical Dispersion: Includes empirical variance and standard deviation.

• Bivariate Analysis: Examines relationships between two variables.

• Correlation Coefficients: Quantifies the strength and direction of relationships.

• Contingency Tables: Displays frequencies of categorical variables.

• Pearson Coefficient: Measures linear correlation between metric variables.

• Ordinal Data: Can be analyzed using rank correlation methods.

7

• Correlation: Describes relationships between two variables using correlation coefficients based on variable types (nominal, ordinal, metric).

• Contingency Tables: Used for two-dimensional frequency distributions; includes conditional frequencies and measures of association.

• Probability Theory: Introduces random experiments, events, and Kolmogorov axioms; covers Laplace experiments and combinatorics.

• Bayes’ Theorem: Explains conditional probability and its application in real-world scenarios, such as medical testing.

• Outcome: Understanding of probability basics, combinatorial calculations, and Bayes’ theorem application.

8

• Random Experiment: Defined by well-defined conditions with unpredictable outcomes (e.g., dice throw).

• Kolmogorov Axioms: Fundamental properties of probability measures.

• Random Variables: Assign outcomes to numbers; can be discrete (countable values) or continuous (any value in an interval).

• Distributions: Includes discrete (e.g., binomial, uniform) and continuous (e.g., normal) distributions.

• Expected Value & Variance: Key metrics for understanding random variables' behavior.

• Applications: Used in statistical tests and linear regression.

9

• Random Variables: Defined as functions mapping outcomes to real numbers.

• Discrete vs. Continuous Distributions: Discrete has countable outcomes; continuous uses probability density functions.

• Simple Linear Regression: Models correlation between independent (X) and dependent (Y) variables.

• Key Concepts:

  • Residual Analysis: Evaluates fit of regression line.

  • Determinacy Measure (R²): Indicates model fit; ranges from 0 to 1.

• Estimation: Parameters (β0, β1) estimated using least squares method.

• Applications: Used in various fields to predict outcomes based on correlations.