R Visualization

Daniel Tafmizi

Dr. Friedman

Lis 4370

Module 9

Github: daniel.R/Work.R/LIS4370Rprog/wineClusterViz.R at main · DanielDataGit/daniel.R

    I did some clustering analysis with the wine dataset that includes Country, alcohol as liters of wine, deaths per 100,000, heart disease per 100,000, and liver disease per 100,000.


Some correlations are prevalent. Unsurprisingly, heart disease and death are positively correlated. Alcohol and liver disease have a positive correlation. Interestingly, heart and alcohol have a negative correlation. 


I chose to use k-means as opposed to knn because of the small dataset. I think the algorithm did a great job of creating clusters. I found that three clusters resulted in the most uniformity.  It is interesting to see where each country ends up on the map.




I thought this was a really cool visualization. It incorporates a dendrogram into the heatmap, showing us the clusters in a different style. This shows us that there are 5 cluster hierarchies. The heatmap aspect allows us to see where the differences reside. The first cluster has high heart and death, and low alcohol and liver. The second is opposite the first. The third is overall average but leans towards death and heart. The fourth sees a spike in heart, but low in others. The fifth has fairly low values for all.


Comments

Post a Comment

Popular posts from this blog

redditApiR

Daniel Tafmizi Dr. Friedman Lis 4370 redditApiR Github:  DanielDataGit/RedditApiR: redditApiRPackage  Authenticating Reddit Account: Create or log into account Ensure account is verified by going to account settings-email-verify Go to  preferences (reddit.com) Create an App (personal use script for data collection needs) Use the following script: library(httr) # Initialize parameters client_id <- "Code under "personal use script" " client_secret <- "next to secret" user_agent <- "app name" username <- "Account username" password <- "Account Password" # Get the access token response <- POST(   "https://www.reddit.com/api/v1/access_token",   authenticate(client_id, client_secret),   body = list(     grant_type = "password",     username = username,     password = password   ),   encode = "form",   user_agent(user_agent) ) # Extract the access token token <- content(response)$acc...
Read more

Mod 7 OOP

Daniel Tafmizi Dr. Friedman Lis 4370 Module 7 1. How do you tell what OO system (S3 vs. S4) an object is associated with? Using the function isS4() or isS3() on the object name will tell you which OO system the object was designed under.  2.  How do you determine the base type (like integer or list) of an object? There are a few useful functions for understanding the storage types of an object. typeof() for storage type, class() for structure type, mode() for data type. 3.What is a generic function? A  function that can be applied to different data and class types. A generic function will have many methods for differing data and class types and call them as needed.  4. What are the main differences between S3 and S4? Many differences lie in how operations are called in either class. These differences arise from the fact that S4 objects require explicit class, structure, and inheritance definitions, while S3 do not. I like S4 objects because they remind of classes in ...
Read more

Description File

Daniel Tafmizi Dr. Friedman Lis 4370 Module 10 Github:  daniel.R/Work.R/LIS4370Rprog/finalproject/DESCRIPTION at main · DanielDataGit/daniel.R     For my final project, I would like to incorporate the fellowship project. In it, I will determine the architecture of my fellowship project, as well as start to compile techniques for ML modelling and its respective visualizations. I plan to define techniques for supervised ML, broken down into classification and regression. I hope to make this as beginner friendly as possible, as I have had and continue to have my fair share of troubles dealing with the ML jargon found in professional sources. Overall, I hope to mimic the scikit-learn python library in a more R/beginner friendly way. If the fellowship project heads in a different direction, this project will still act as a great step in understanding how I can frame different elements for ML visualization.
Read more