Lecture 22

class: title-slide, left, bottom

# Lecture 22
----
## **DANL 200: Introduction to Data Analytics**
### Byeong-Hak Choe
### November 17, 2022

---
# Announcement
### <p style="color:#00449E"> Office Hours

- On November 21, Monday, I will have Zoom office hours from 3:30 PM to 5:30 PM.

---
# For Your Information
### <p style="color:#00449E"> Google Data Analytics Certificate
 
Course programs for [Google Data Analytics Certificate](https://grow.google/certificates/data-analytics/#?modal_active=none) are hosted on [Coursera](https://www.coursera.org/professional-certificates/google-data-analytics?utm_source=google&utm_medium=institutions&utm_campaign=gwgsite&_ga=2.19327484.536055472.1668546492-1708405686.1662938381).

- Foundations: Data, Data, Everywhere
- Ask Questions to Make Data-Driven Decisions
- Prepare Data For Exploration
- Process Data from Dirty to Clean
- Analyze Data to Answer Questions
- Share Data Through the Art of Visualization
- Data Analysis with R Programming
- Data Analytics Capstone Project: Complete a Case Study

---
# For Your Information
### <p style="color:#00449E"> Google Data Analytics Certificate

- Google Data Analytics Certificate program uses *R*.

- R is a great starting point for foundational data analysis, and offers helpful packages for beginners to apply to their projects.

- The data analysis tools and platforms included in the certificate curriculum are spreadsheets (Google Sheets or Microsoft Excel), SQL, presentation tools (Powerpoint or Google Slides), Tableau, RStudio, and Kaggle.

---
# For Your Information
### <p style="color:#00449E"> Why Python, R, and SQL?

---
# For Your Information
### <p style="color:#00449E">  Stack Overflow Trends

- We can see the trends of programming based on tags on the website, [Stack Overflow](https://insights.stackoverflow.com/trends?tags=r%2Cpython%2Csql)

---
class: inverse, center, middle

# Modeling Methods - Linear Regression
<html><div style='float:left'></div><hr color='#EB811B' size=1px width=796px></html>

---
# Linear Regression 
### <p style="color:#00449E">  Example

- Suppose we also want to estimate how gender will affect personal income. 
- Linear regression assumes that ...
  - The outcome `PINCP[i]` is linearly related to each of the inputs `AGEP[i]` and `SEX[i]`:

`$$\texttt{PINCP[i]} \;=\quad \texttt{f(AGEP[i], SEX[i])} \,+\, \texttt{e[i]} \qquad\qquad\qquad\qquad\\
\;=\quad \texttt{b0} \,+\, \texttt{b1*AGEP[i]} \,+\, \texttt{b2*SEX[i]}\,+\, \texttt{e[i]}$$`
- A variable on the left-hand side is called an outcome variable or a dependent variable.
- Variables on the right-hand side are called explanatory variables, independent variables, or input variables.
- Coefficients `$\texttt{b[1]}, ... , \texttt{b[P]}\;$`  on the right-hand side are called beta coefficients.

---
# Linear Regression 
### <p style="color:#00449E">  Goals of Linear Regression

-  The goals of linear regression are  ... 
  1. Find the estimated values of `b1` and `b2`: `$\quad \hat{\texttt{b1}}$` and `$\hat{\texttt{b2}}$`.
  
  2. Make a prediction on `PINCP[i]` for each person `i`: `$\quad \widehat{\texttt{PINCP}}\texttt{[i]}$`.

`$$\widehat{\texttt{PINCP}}\texttt{[i]} \;=\quad \hat{\texttt{b0}} \,+\, \hat{\texttt{b1}}\texttt{*AGEP[i]} \,+\, \hat{\texttt{b2}}\texttt{*SEX[i]}$$`

- We will use the hat notation `$(\,\hat{\texttt{ }}\,)$`  to distinguish *estimated* beta coefficients and *predicted* outcomes from *true* values of beta coefficients and *true* values of outcome variables, respectively.

---
# Linear Regression 
### <p style="color:#00449E">  More Assumptions

- Assumptions on the linear regression model are that ...

- The outcome variable is a linear combination of the explanatory variables.
  
  - Errors have a mean value of 0.
  
  - Errors are *uncorrelated* with explanatory variables.

---
# Linear Regression 
### <p style="color:#00449E"> Beta estimates

- Linear regression finds the beta coefficients `$( \texttt{b[0]}, ... , \texttt{b[P]} )$` such that ...

– The linear function `$\texttt{f(x[i, ])}$` is as near as possible to
`$\texttt{y[i]}$` for all `$\texttt{(x[i, ], y[i])}$` pairs in the data.

- In other words, the estimator for the beta coefficients is chosen to minimize the sum of squares of the *residual errors* (SSR):

-  `$\texttt{Residual_Error[i] = y[i] - } \hat{\texttt{y}}\,\texttt{[i]}$`.
  
  -  `$\texttt{SSR} = \texttt{Residual_Error[1]}^{2} + \cdots + \texttt{Residual_Error[N]}^{2}$`.

---
# Linear Regression 
### <p style="color:#00449E"> Example of Prediction

.pull-left[
- Linear regression often does an excellent job, even when the actual relation between `$\texttt{x[i, ]}$` and `$\texttt{y[i]}$` is not linear.

- For example, `$y = x^2\;$` vs. `$\;f(x) = -22 + 11x$`

]

.pull-right[

]

---
# Linear Regression 
### <p style="color:#00449E"> Evaluating Models

- **Training data**: When we're building a model to make predictions or to identify the relationships, we need *data* to build the model.

- **Testing data**: We also need data to test whether the model works well on *new data*.

.pull-left[

- So, we split data into training and test sets when building a linear regression model.
]

.pull-right[
<img src="../lec_figs/pds_fig4_12.png" width="100%" style="display: block; margin: auto;" />
]

---
# Linear Regression 
### <p style="color:#00449E"> Evaluating Models
- We need to ensure that our model will perform well in the real world.

<div class="figure" style="text-align: center">
<img src="../lec_figs/pds_fig6_6.png" alt="Schematic of model construction and evaluation" width="52%" />
<p class="caption">Schematic of model construction and evaluation</p>
</div>

---
# Linear Regression 
### <p style="color:#00449E"> A Little Bit of Statistics for the Uniform Distribution

.pull-left[

- The probability density function for the uniform distribution looks like:

- With the uniform distribution, any values of `$x$` between 0 and 1 is equally likely drawn.

- `runif(n)` generates `n` uniform random numbers between 0 and 1.

]

.pull-right[

]
- We will use the uniform distribution when splitting data into training and testing data sets.

---
class: inverse, center, middle

# Linear Regression using **R**
<html><div style='float:left'></div><hr color='#EB811B' size=1px width=796px></html>

---
# Linear Regression 
### <p style="color:#00449E"> Example of Linear Regression using **R**

- We will use the 2016 US Census PUMS dataset.
  - Full-time employees between 20 and 50 years of age with income between $1,000 and $250,000;

- Personal data recorded includes personal income and demographic variables:
  - `PINCP`: personal income
  - `AGEP`: age 
  - `SEX`: sex

---
# Linear Regression 
### <p style="color:#00449E"> Spliting Data into Training and Testing Data

.panelset[

.panel[.panel-name[Step 1. set.seed()]

```r
# Importing the cleaned small sample of data
psub <- readRDS( url('https://bcdanl.github.io/data/psub.RDS') )

# Making the random sampling reproducible by setting the random seed.
set.seed(3454351) # 3454351 is just any number.
# The set.seed() function sets the starting number 
# used to generate a sequence of random numbers.

# With set.seed(), we can replicate the random number generation:
# If we start with that same seed number in the set.seed() each time, 
# we run the same random process, 
# so that we can replicate the same random numbers.
```

]

.panel[.panel-name[Step 2. runif()]

```r
# How many random numbers do we need?
gp <- runif([?]) 
# a number generation from a random variable that follows Unif(0,1)

# Splits 50-50 into training and test sets 
# using filter() and gp

dtrain <- filter(psub, gp >= [?]) 
dtest <- filter(psub,  gp < [?])
# A vector can be used for CONDITION in the filter(data.frame, CONDITION) 
# if the length of the vector is the same as that of the data.frame.
```
]

]

---
# Linear Regression 
### <p style="color:#00449E"> Exploratory Data Analysis (EDA)

- Use summary statistics and visualization to explore the data, particularly for the following variables:

- `PINCP`: personal income
  - `AGEP`: age 
  - `SEX`: sex

- It's often a better idea to get some sense of how the data behaves through EDA before doing any statistical analysis.

```r
# install.packages("GGally")  # to use GGally::ggpairs()
ggpairs( select(dtrain, PINCP, AGEP, SEX) )  # for correlogram or correlation matrix
```

---
# Linear Regression 
### <p style="color:#00449E"> Building a linear regression model using `lm()`

```r
model <- lm(formula = PINCP ~ AGEP + SEX, 
            data = dtrain)
```
In the above line of R commands, ...
- `model`: R object to save the estimation result of linear regression
-  `lm()`: Linear regression modeling function
- `PINCP ~ AGEP + SEX`:  Formula for linear regression
- `PINCP`: Outcome/Dependent variable
- `AGEP, SEX`: Input/Independent/Explanatory variables
- `dtrain`: Data frame to use for training

---
# Linear Regression using **R**
### <p style="color:#00449E"> Making predictions with a linear regression model using `predict()`

```r
dtest$pred <-  predict(model, 
                       newdata = dtest)
```

- In the above line of R commands, ...
  - `dtest$pred`: Adding a new column `pred` to the `dtest` data frame. `mutate()` also works.
  - `predict()`: Function to get the predicted outcome using `model` and `dtest`
  - `model`: R object to save the estimation result of linear regression
  - `dtest`: Data frame to use in prediction

- We can make prediction using `dtrain` data frame too.

---
# Linear Regression using **R**
### <p style="color:#00449E"> Summary of the regression result

```r
summary(model)   # This produces the output of the linear regression.
```

---
# Linear Regression using **R**
### <p style="color:#00449E"> Indicator variables

- Linear regression handles a factor variable with `m` possible levels by converting it to `m-1` indicator variables, and the rest `1` category, the first level of the factor variable, becomes a reference level.

- The value of any indicator variable is either 0 or 1.

- E.g., the indicator variable, `SEXFemale`, is follows:

$$
\texttt{SEXFemale[i] }\\
= \begin{cases}
\texttt{1} & \text{if a person } \texttt{i} \text{ is } \texttt{female};\\\\
\texttt{0} & \text{otherwise}.\qquad\qquad\quad\,
\end{cases}
$$

- The level `male` becomes a reference level when interpreting the beta estimate for `SEXFemale`.

---
# Linear Regression using **R**
### <p style="color:#00449E"> Setting a reference level

- If the independent variable includes factor variables, we can set a reference level for each factor variable using `relevel(VARIABLE, ref = "LEVEL")`.

.panelset[

.panel[.panel-name[code]

```r
dtrain$SEX, <- relevel(dtrain$SEX, ref = "Male")

model <- lm(log(PINCP) ~ AGEP + SEX, 
            data = dtrain)
            
summary(model)
```
]

.panel[.panel-name[variable]

- E.g., the indicator variable, `SEXMale`, is follows:

$$
\texttt{SEXMale[i] }\\
= \begin{cases}
\texttt{1} & \text{if a person } \texttt{i} \text{ is } \texttt{male};\\\\ 
\texttt{0} & \text{otherwise}.\qquad\qquad\quad
\end{cases}
$$

- The level `female` now becomes a reference level.

- Note: Changing the reference level does not change the regression result.
]

]

---
# Linear Regression using **R**
### <p style="color:#00449E"> Interpreting Estimated Coefficients
The model is ...

`$$\texttt{PINCP[i]} \;=\quad \texttt{b0} \,+\, \texttt{b1*AGEP[i]} \,+\,\texttt{b2*SEX.Male[i]}\,+\, \texttt{e[i]}$$`

All else being equal, ...

.panelset[

.panel[.panel-name[`AGEP`]

- All else being equal, an increase in `AGEP` by one unit is associated with an increase in `PINCP` by `b1`.

]

.panel[.panel-name[`SEX.Male`]

- All else being equal, an increase in `SEX.Male` by one unit is associated with an increase in `PINCP` by `b2`.

- All else being equal, being a male relative to being a female is associated with an increase in `PINCP` by `b2`.

]

---
# Linear Regression using **R**
### <p style="color:#00449E"> Interpreting Estimated Coefficients
Consider the predicted incomes of the two male persons, `Ben` and `Bob`, whose ages are 51 and 50 respectively.

`$$\widehat{\texttt{PINCP[Ben]}} \;=\quad \hat{\texttt{b0}} \,+\, \hat{\texttt{b1}}\texttt{ * AGEP[Ben]} \,+\, \hat{\texttt{b2}}\texttt{ * SEX.Male[Ben]}\\
\widehat{\texttt{PINCP[Bob]}} \;=\quad \hat{\texttt{b0}} \,+\, \hat{\texttt{b1}}\texttt{ * AGEP[Bob]} \,+\, \hat{\texttt{b2}}\texttt{ * SEX.Male[Bob]}$$`

`$$\Leftrightarrow\qquad\widehat{\texttt{PINCP[Ben]}} \,-\, \widehat{\texttt{PINCP[Bob]}}\qquad  \\
\;=\quad \hat{\texttt{b1}}\texttt{ * }(\texttt{AGEP[Ben]} - \texttt{AGEP[Bob]})\\
\;=\quad \hat{\texttt{b1}}\texttt{ * }\texttt{(51 - 50)}\qquad\qquad\quad\;\;\\
\;=\quad \hat{\texttt{b1}}\qquad\qquad\qquad\qquad\quad\;\;\;\,$$`

---
# Linear Regression using **R**
### <p style="color:#00449E"> Interpreting Estimated Coefficients
Consider the predicted incomes of the two persons, `Ben` and `Linda`, whose ages are the same as 50. `Ben` is `male` and `Linda` is `female`.

`$$\widehat{\texttt{PINCP[Ben]}} \;=\quad \hat{\texttt{b0}} \,+\, \hat{\texttt{b1}}\texttt{ * AGEP[Ben]} \,+\, \hat{\texttt{b2}}\texttt{ * SEX.Male[Ben]}\;\;\,\\
\widehat{\texttt{PINCP[Linda]}} \;=\quad \hat{\texttt{b0}} \,+\, \hat{\texttt{b1}}\texttt{ * AGEP[Linda]} \,+\, \hat{\texttt{b2}}\texttt{ * SEX.Male[Linda]}$$`

`$$\Leftrightarrow\qquad\widehat{\texttt{PINCP[Ben]}} \,-\, \widehat{\texttt{PINCP[Linda]}}\qquad\qquad\qquad  \\
\;=\quad \hat{\texttt{b2}}\texttt{ * }(\texttt{SEX.Male[Ben]} - \texttt{SEX.Male[Linda]})\\
\;=\quad \hat{\texttt{b2}}\texttt{ * }\texttt{(1 - 0)}\qquad\qquad\quad\qquad\qquad\quad\;\;\\
\;=\quad \hat{\texttt{b2}}\qquad\qquad\qquad\qquad\quad\qquad\qquad\quad\;$$`

---
# Linear Regression using **R**
### <p style="color:#00449E"> Interpreting Estimated Coefficients

- What does it mean for a beta estimate `$\hat{\texttt{b}}$` to be statistically significant at 5% level?

- It means that the null hypothesis `$H_{0}: \texttt{b} = 0$` is rejected for a given significance level 5%.
  
  - "2 standard error rule" of thumb: The true value of `$\texttt{b}$` is 95% likely to be in the confidence interval `$(\, \hat{\texttt{b}} - 2 * \texttt{Std. Error}\;,\; \hat{\texttt{b}} + 2 * \texttt{Std. Error} \,)$`.
  
  - The standard error tells us how uncertain our estimate of the coefficient `b` is.
 
  - We should look for the stars!

---
# Linear Regression using **R**
### <p style="color:#00449E"> **R-squared**

- **R-squared** is a measure of how well the model “fits” the data, or its “goodness of fit.”
  - **R-squared** can be thought of as *what fraction of the `y`'s variation is explained by the independent variables*.
  
  
- **R-squared** will be higher for models with more explanatory variables, regardless of whether the additional explanatory variables actually improve the model or not.
- We want **R-squared** to be *fairly* large and **R-squareds** that are similar on testing and training.

- The adjusted **R-squared** is the multiple **R-squared** penalized for the number of input variables.

---
# Linear Regression using **R**
### <p style="color:#00449E"> Visualizations to diagnose the quality of modeling results

.panelset[
.panel[.panel-name[Quality?]
- The following two visualizations from the linear regression are useful to determine the quality of linear regression:

1. Actual vs. predicted outcome plot;
  2. Residual plot.

]

.panel[.panel-name[Actual vs. Predicted]

- The following is **the actual vs predicted outcome plot**.

```r
ggplot( data = dtest, 
        aes(x = pred, y = PINCP) ) +
  geom_point( alpha = 0.2, color = "darkgray" ) +
  geom_smooth( color = "darkblue" ) +  
  geom_abline( color = "red", linetype = 2 )  # y = x, perfect prediction line
```
]

.panel[.panel-name[Residuals]

- The following is **the residual plot**.

- `Residual[i] = y[i] -  Predicted_y[i]`.

```r
ggplot(data = dtest, 
       aes(x = pred, y = PINCP - pred)) +
  geom_point(alpha = 0.2, color = "darkgray") +
  geom_smooth( color = "darkblue" ) +   
  geom_hline( aes( yintercept = 0 ),  # perfect prediction 
              color = "red", linetype = 2) + 
  labs(x = 'Predicted PINCP', y = "Residual error")
```

]

.panel[.panel-name[Q & A]

- From the plot of actual vs. predicted outcomes and the plot of residuals, we should ask the following two questions ourselves:

- On average, are the predictions correct?
  - Are there systematic errors?

- A well-behaved plot will bounce *randomly* and form a cloud roughly around the perfect prediction line.

]

.panel[.panel-name[Sys. Errors]
.pull-left[
<img src="../lec_figs/pds_fig7_8.png" width="100%" style="display: block; margin: auto;" />
]
.pull-right[
<div class="figure" style="text-align: center">
<img src="../lec_figs/residual-hetero.png" alt="An example of systematic errors in model predictions" width="90%" />
<p class="caption">An example of systematic errors in model predictions</p>
</div>
]

]

---
# Linear Regression using **R**
### <p style="color:#00449E"> Practical considerations in linear regression

.panelset[
.panel[.panel-name[Correlation vs. Causation]

- Correlation does not imply causation:

- Just because a coefficient is significant, doesn’t mean our explanatory variable causes the response of our outcome variable.
  
  - In order to test cause-and-effect relationships through regression, we would often need data from (quasi-)experiments to remove *selection bias*.

]

.panel[.panel-name[RCT & A/B testing]

- To achieve causality, researchers conduct experiments such as randomized controlled trials (RCT) and A/B testing:
  - The treatment group receives the treatment whose effect the researcher is interested in.
  - The control group receives either no treatment or a placebo. 
  - The treatment variable indicates the status of treatment and control.
 
 
- In linear regression, if all explanatory variables apart from the treatment variable are made equal across the two groups, selection bias is mostly eliminated, so that we may infer *causality* from beta estimates. 
  
]

.panel[.panel-name[Practical significance]
- There is a difference between practical significance and statistical significance:

- Whether an association between `x` and `y` is *practically significant* depends heavily on *the unit of measurement*.
  
  - E.g., We regressed income (measured in $) on height, and got a statistically significant beta estimate of 100, with a standard error of 20. 
  
  - **Q**. Is 100 a large effect?
]
]

---
# Linear Regression using **R**
### <p style="color:#00449E"> More Explanatory Variables in the Model

- In the 2016 US Census PUMS dataset, personal data recorded includes occupation, level of education, personal income, and many other demographic variables:

- `COW`: class of worker
  - `SCHL`: level of education

---
# Linear Regression using **R**
### <p style="color:#00449E"> More Explanatory Variables in the Model

- Suppose we also want to assess how personal income varies with (1) a bachelor's degree, (2) a class of work on personal income, (3) age, and (4) gender.

1. Conduct the exploratory data analysis.
  2. Based on the visualization, set a hypothesis regarding the relationship between having bachelor's degree and `PINCP`.
  3. Train the linear regression model.
  4. Interpret the beta coefficients from the linear regression result.
  5. Calculate the predicted `PINCP` using the testing data.
  6. Draw the actual vs. predicted outcome plot and the residual plot.