EPPS 6356 Assignment 10

EPPS 6356 Assignment 10: Data Analytics Colloquium Review

Write a review on visualization methods used in the presentation by the speaker and post on own GitHub website.

Here is a link to the Colloquium: https://youtu.be/q-37SPXg7ko

Here is my review of the Colloquium:

I am going to be completely 100 percent candid: this colloquium was very challenging and difficult for me to understand. This is my very first semester intensively using R Studio and coding statistical analyses, let alone actually analyzing those statistical visualizations that I produce! Furthermore, I normally only am required to use generic regressions for my other courses; we really never really need to use other data visualization methods in order to demonstrate or display our important statistical findings. The fanciest thing I have ever done in R Studio prior to this course was an interaction effects plot, which honestly, I still have issues interpreting! All this is being said in order to give you an idea of my background knowledge of statistical methods, visualization, and interpretation skills: they are very minimal. Here is what I understood from the colloquium.

Dr. Patrick Brandt discussed multiple categories or “types” or models. These models were specific to “Statistics & Time Series for Policy Intervention & Change Identification”, which was the topic of the colloquium presentation. There were four types of models or visualizations that Dr. Brandt described: Type 0 (Basic Time Series), Type 1 (Binary Segmentation), Type 2 (Regularization and Fused Lasso), and Type 3 (Bayesian methods). Below are further descriptions of each method of these data visualizations:

Model Type 0: Basic Time Series Model

• Simple model where the moment where the intervention occurs is identifiable.

• Easy for estimations.

• Models allow for many differences; Type 0 does not necessarily show distinct difference between multiple factors.

• Example used: Gas prices and seat belts; Salmonella in chicken breeding

Model Type 1: Binary Segmentation

• This model is commonly used for trends that are more complex, deterministic, or stochastic.

• What is stochastic? According to the Oxford Dictionary, “stochastic” means that the data/values are “randomly determined; having a random probability distribution or pattern that may be analyzed statistically but may not be predicted precisely”.

• Example used: US Industrial Revolution and COVID.

Model Type 2: L1/L0 Regularization and Fused Lasso

• This type of model treats the issue by identifying a correct set of covariates from the possible sets of interventions presented by the data.

• There are different ways to identify these covariates:

  • Search for each of the change-points individually.

  • Assessing the probability of a change-point.

  • Fused Lasso Model: univariate vs. multivariate analyses

  • (Safikhani et al. 2022).

Model Type 3: Bayesian Methods

• This method was mentioned briefly, but not discussed in detail.

• Here is a definition of Bayesian models from a quick Google search: “A Bayesian model is a statistical model where you use probability to represent all uncertainty within the model, both the uncertainty regarding the output but also the uncertainty regarding the input (aka parameters) to the model.”

# Type 0: Times Series Model

lapply(c("quantmod", "tidyverse","TSstudio"), require, character.only = TRUE)

Loading required package: quantmod

Loading required package: xts

Loading required package: zoo

Attaching package: 'zoo'

The following objects are masked from 'package:base':

    as.Date, as.Date.numeric

Loading required package: TTR

Registered S3 method overwritten by 'quantmod':
  method            from
  as.zoo.data.frame zoo 

Loading required package: tidyverse

── Attaching packages ─────────────────────────────────────── tidyverse 1.3.2 ──
✔ ggplot2 3.4.0      ✔ purrr   0.3.4 
✔ tibble  3.1.6      ✔ dplyr   1.0.10
✔ tidyr   1.2.0      ✔ stringr 1.4.0 
✔ readr   2.1.1      ✔ forcats 0.5.1 
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::first()  masks xts::first()
✖ dplyr::lag()    masks stats::lag()
✖ dplyr::last()   masks xts::last()
Loading required package: TSstudio

[[1]]
[1] TRUE

[[2]]
[1] TRUE

[[3]]
[1] TRUE

library(TSstudio)
quantmod::getSymbols("TSLA", src="yahoo")

[1] "TSLA"

ts_plot(TSLA$TSLA.Adjusted, 
        title = "Tesla Stock prices",
        Ytitle = "")