Statistics for data science
Download as PPTX, PDF10 likes8,972 views
The document provides an overview of statistics for data science. It introduces key concepts including descriptive versus inferential statistics, different types of variables and data, probability distributions, and statistical analysis methods. Descriptive statistics are used to describe data through measures of central tendency, variability, and visualization techniques. Inferential statistics enable drawing conclusions about populations from samples using hypothesis testing, confidence intervals, and regression analysis.
Statistics for data science
- 1. zekeLabs Statistics for Data Science
- 2. “Goal - Become a Data Scientist” “A Dream becomes a Goal when action is taken towards its achievement” - Bo Bennett “The Plan” “A Goal without a Plan is just a wish”
- 3. ● Introduction to Statistics ● Importance of Statistics ● Understanding Variables Types ● Descriptive vs Inferential Statistics Overview of Statistics
- 4. Introduction to Statistics ● Science of learning from data. ● Methodical data collection. ● Employ correct data analysis. ● Presenting analysis effectively. ● Opposite to statistics is “Anecdotal Evidence”.
- 5. Importance ● Avoid getting biased samples ● Prevent overgeneralization ● Wrong causality ● Incorrect Analysis ● Applied to any domain
- 6. Variables ● Explanatory (predictor or independent) ● Response (outcome or dependent) ● A variable can serve as independent in one study and dependent in another
- 7. Data Types of Variables - Quantitative versus Qualitative ● Quantitative - Numerical data. Eg. weight, temperature, number_project ● Qualitative - Non-numerical data. Eg. dept, salary
- 8. Types of Quantitative Variables ● Continues - any numeric value. Eg. Sqft ● Discrete - count of the presence of a characteristic, result, item, or activity. Eg. Floor
- 9. Qualitative Data: Categorical, Binary, and Ordinal ● Categorical or Nominal. Eg - dept ( sales, RD etc. ) ● Binary. Eg. Left ( 1 or 0 ) ● Ordinal. Eg. salary ( low, medium, high )
- 10. Choosing Statistical Analysis based on data type
- 11. Types of Statistical Analysis ● Descriptive Statistics - Describes data. ○ Common Tools - Central tendency, Data distribution, skewness ● Inferential Statistics - Draw conclusions from the sample & generalize for entire population ○ Common Tools - Hypothesis Testing, Confidence Intervals, Regression Analysis
- 12. ● Measure of Central Tendency ● Measure of Variability ● Visualizing Data Summarizing Data
- 13. Measure of Central Tendency ● Mean - Average of data, suited for continuous data with no outliers ● Median - Middle value of ordered data, suited for continuous data with outliers ● Mode - Most occuring data, suited for categorical data ( both nominal and ordinal )
- 14. Measure of Variance ● Range ● Interquartile Range ● Variance ● Standard Deviation
- 15. Visualizing Continuous Data ● Histogram ● ScatterPlot
- 16. Visualizing Continuous Data - 2 ● Box-Plot
- 17. Visualizing Discrete Data ● Histogram ● Pie
- 18. ● Basics of Probability ● Conditional Probability ● Discrete Probability Function ● Continuous Probability Function ● Central Limit Theorem Probability Distribution
- 19. Probability of Single Event
- 20. Probability of Two Independent Events P(A AND B) = P(A) * P(B) Probability of heads on tossing of two coins P(A) * P(B) = ½ * ½ = ¼ P(A OR B) = P(A) + P(B) - P(A AND B) Probability of head in 1st flip or probability of head in 2nd flip or both ½ + ½ - ¼ = ¾
- 21. Conditional Probability Probability of an event given the other event has occurred. P(B|A) - Probability of event B given A has happened P(A AND B) = P(A) * P(B|A) Probability of drawing 2 aces = P(drawing one ace from deck) * P(drawing one ace given already one ace is pulled out) Probability of drawing 2 aces = 4/52 * 3/51
- 22. Probability distribution ● A function describing the likelihood of obtaining possible values that a random variable can assume. ● Consider salary of employee data, we can create distribution of salary. ● Such distribution is useful to know which outcome is more likely. ● Sum of probability of all outcomes is 1, so every outcome has likelihood between 0 & 1 ● PDF are divided into two types based on data - Discrete and Continues
- 23. Discrete Probability Distribution Function ● Probability mass functions for discrete data ● Binomial Distribution for Binary Data (Yes/No) ● Poisson Distribution for count data (No. of cars per family) ● Uniform Distribution for Data with equal probability (Rolling dice)
- 27. Probability distribution for continuous data ● Probability mass function for continuous data ● Central tendency, variation & skewness important parameters ● Normal Probability Distribution or Gaussian Distribution or Bell curve ● Lognormal Probability Distribution
- 28. Normal Distribution ● A probability function that describes how the values of a variable are distributed. ● Symmetric distribution ● Mean = 69, Std = 2.8 ● Notation Alert, mu & sigma term used for entire population Height Distribution
- 29. Normal Distribution - 2 ● Empirical Rule of Normal Distribution : 68 - 95 - 99 ● Standard Normal Distribution : Mean = 0, Std = 1.0 ● Z-scores is a great way to understand where a specific observation fall wrt entire population. It is basically number of std far from mean.
- 31. ● Introduction ● Central Tendency ● Data Distribution ● Skewness ● Correlation Descriptive Statistics
- 33. ● Introduction ● Hypothesis Testing ● Confidence Intervals ● Regression Analysis Inferential Statistics
- 34. ● Chi-square Test of Independence ● Correlation and Linear Regression ● Analysis of Variance or ANOVA Relationships between Variables
- 35. Thank You !!!
- 36. Visit : www.zekeLabs.com for more details Let us know how can we help your organization to Upskill the employees to stay updated in the ever-evolving IT Industry. www.zekeLabs.com | +91-8095465880 | [email protected]