Course Description
Do you have these questions playing on your mind?
Should I open a new store for my business? Should I launch new product? Should I renew contract of my vendor? Whom should I give more bonuses among my team members? Who will win cricket world cup? Who will win elections?
AND how many times have you gone wrong with such decision making?
- Answer will be an obvious – “I was wrong Many Times” if the decision making is based on intuition.
- Answer will be an obvious – “I was right Many Times” if the decision making is based on statistical analysis.
- Statistical analysis is used to take decisions across industries and across job levels. Intuition based decision making should be coupled with a factor called LUCK, for success. Most companies have not even seen the light because of lack of statistical analysis. We at ExcelR expose you to various statistical techniques for data analysis using a tool called “R/RStudio”. Revolution Analytics was recently acquired by Microsoft but will still continue to be an open source statistical software. With the tremendous increase in data generation which is set to increase by 4 thousand times by 2020 brings with it a new challenge on how to analyze the data faster to give business leaders insights for decision making. This is where statistical analysis comes in handy which forms the spine strength for any data scientist. Rather than trusting the gut feel, a data scientist uses statistical data analysis to bring in business value.
Course Curriculum
- Motivate the use of statistical methods for managerial decision making
- Discuss the concepts of probability distributions and random variables
- Review methods of representing data, pictorially and through summary statistics
- Introduce standard normal distribution
- Discuss applications of normal distribution
- Introduce the concept of statistical inference
- Recognize the existence of sample-to-sample variations
- Understand central limit theorem and its implications for statistical inference
- Introduce the concept of confidence intervals as a way to make statistical inferences
- Calculate confidence intervals for population mean with known and unknown population
- Calculate confidence intervals for population proportions
- Calculate confidence intervals for population variance
- Quantify minimum sample sizes to achieve certain margin of error in predictions
- Learn how to state null and alternative hypotheses
- Understand type-I and type-II errors
- Conduct one-sided hypothesis test for population proportion / mean
- Conduct two-sided hypothesis tests for population proportion / mean
- Compare the means using paired observations
- Test for the difference of two population means using independent samples
- Test for the difference of two population proportions
- Introduce Design of Experiments
- Conduct one way Analysis of Variance (ANOVA)
- Introduce the notion of statistical tests on ordinal data
- Test for the difference between mean ranks using paired observations
- Compare mean ranks in two independent samples
- Bivariate data
- Scatter plot
- Compare mean ranks in two independent samples
- Covariance
- Correlation coefficient
- Uses and issues
- Linear regression
- Covariance
- Assumptions
- Scatter plot matrix
- Multiple linear regression
- Ordinary Least Squares method (OLS)
- Basic regression summary
- Interpretation of coefficient estimates
- standard errors
- t-values and p-values
- Basic tests
- ANOVA table
- Generalized Linear Models
- Binary and multinomial logistic regressions
- Poisson regression
- Zero-inflated Poisson regression
- Negative Binomial regression
- Missing value patterns: Missing completely at random (MCAR). Missing at random (MAR). Missing not at random (MNAR)
- List wise deletion. Pairwise deletion
- Various imputation methods: Hot deck imputation. Mean substitution. Regression imputation. EM imputation
- Censoring and truncation. Characteristics of survival analysis data
- Time-to-event data. Hazard and survival functions
- Kaplan-Meier estimate of survival function–
- Cox proportional hazards model (ph), estimation and its analysis. Extensions
- Stratified ph; ph with time-varying covariates–
- Parametric survival analysis with standard distributions
- Accelerated failure time models
- Basic concepts: randomization, replication and control
- Experimental design for testing differences in several means: Completely randomized and randomized–complete block designs. Cross-over designs
- Two-level factorial experiments full and fractional. Plackett-Burman designs
- Designs for three or more levels. Taguchi designs. Response surface designs
- Case-Control designs for campaign evaluation
- Designs for conjoint analysis
Contact Our Team of Experts
