Data Analytics for Cyber Security Defense
November 10, 2016
Fri, Nov 11
Data Analytics for Cyber Security
- Defining Big Data Analytics
- Data types
- Analysis types
- Visualization
- Why use visualization
- See patterns
- Enable participation
- Simple Data Visualization for Distributions
- Histograms
- Illustrate the shape of data distribution
- Example: Exam Results (score frequency distribution)
- Example: Solar production by day and by hour
- Demonstrate with real estate data (Google Sheets)
- Boxplots
- What are Boxplots
- Five number graphic summary: 25th, 50th, 75th Percentile + Min and Max value
- Example: Stacking Bar Charts, Histograms and Boxplots
- Example: Boxplot paired with Histogram
- Histograms
- Why use visualization
- Correlation and Simple Regression
- Scatter plots
- Colorado TCAP Scores and % Free-Reduced Lunch
- Gender Wage Gap
- Bubble Charts
- Animated Bubble Chart
- Correlation
- Pearson Correlation Coefficient
- Correlation Coefficient Matrix
- Spurious Correlation
- Correlation does not imply Causation
- Linear Regression
- Simple (or Bivariate) regression
- Linear, Polynomial equation
- R-squared
- Multiple Regression with XL Miner
- Scatter plots
- Probability for Normal Distributions
- Normal (Bell-shaped) Curve
- Standard Normal and Z-Scores
- Standard Normal Distribution
- Mean = 0
- Standard Deviation = 1
- N(0, 1)
- Z-Scores
- Z = (x – mean)/std dev
- Example
- Young women heights (inches)
- N(64.5, 2.5)
- Z-Score for woman 68 inches tall; Z = (68 – 64.5)/2.5 = 1.4
- Z-Score for woman 60 inches tall; Z = (60 – 64.5)/2.5 = -1.8
- Z-table
- Empirical Rule Probabilities- 68-95-99.7
- Normal Curve Calculations with Z-Scores
- Standard Normal Distribution
- Why Normal Distribution is the most important distribution
- Statistical Computing
- Key technologies: R, Python are in, SAS, SPSS, Minitab et al are out
- Python vs R
- Python
- Numpy, Scipy, etc
- PythonAnywhere.com
- Repl.it
- Wrap-up
2 Comments
Great class, I really enjoyed and benefited from the time with Justin.
I enjoyed it too, Bill! Thanks for a great session. Best, Justin