ny time article tracking location data

https://www.nytimes.com/interactive/2019/12/19/opinion/location-tracking-cell-phone.html

Rename all files in a folder with PowerShell

dir | rename-item -NewName {$_.name -replace ".xlsx", "-copy.xlsx"}

Excel : vlookup to ..lookup

How to lookup if a value in a given cell is present in a set of values?

Using VLOOKUP

=VLOOKUP(A1;H:H;1;FALSE)

R - Importing an Excel file

The R-library openxlsx offers a performant tool for importing Excel files.

Example for openxlsx::read.xlsx:

library(openxlsx)
openxlsx::read.xlsx(file, 
  sheet = 1, 
  colNames = T, 
  rows = seq(2,100)

This package performs better than the read.xlsx in package xlsx. However, there is one thing to notice: columns in date format will not be correctly handled by openxlsx::read.xlsx. The function xlsx::read.xlsx handles them correctly.

Example for xlsx::read.xlsx:

library(xlsx)
xlsx::read.xlsx(file,
  sheetIndex = 1,
  header = T, 
  rowIndex = seq(2,100)

There is also read.xlsx2 function in the library xlsx, which is written in java, and performs better than xlsx::read.xlsx.

library(xlsx)
xlsx::read.xlsx2(file,
  sheetIndex = 1,
  header = T, 
  rowIndex = seq(2,100)

Using the xslx functions requires that the environments is aware of the location of the java runtime:

Sys.setenv(JAVA_HOME='C:\\ieu\\java\\openjdk-11') 

R: create a list of dataframes

gnr1 <- c("1111","4444", "3333","5555", "2222","9999")
gnr2 <- c("7777", "2222","1111","5555","3333","4444","8888")
prime1 <- c(1000, 4000, 3000, 5000, 2000, 9999)
prime2 <- c(7777, 2001, 1001, 5001, 3001, 4001,8888)
val1 <- c(1,2,3,4,5,6)
df1 <- data.frame(gnr1)
df2 <- data.frame(gnr2)
df1$prime <- prime1
df2$prime <- prime2
df1$val <- val1
mylist <- list()
mylist[[1]] <- df1
mylist[[2]] <- df2

R: DataFrames : creating a R Dataframe

# how to create a dataframe in r > diets <- data.frame ('diet'=1:4, 'protein'=c(0,0,1,1), 'vitamin'=c(0,1,0,1))

R: DataFrame: Selecting A Subset of a R Data Frame, merging.

Notes from the Tutorial "Meet The R Dataframe: Examples of Manipulating Data In R":

Use the ChickWeight dataset for this example

data("ChickWeight")


> head (ChickWeight)
  weight Time Chick Diet
1     42    0     1    1
2     51    2     1    1
3     59    4     1    1

Selecting A Subset of a R Data Frame
1. using the function subset
> subset(ChickWeight, Diet==4)

2. with a conditional indexing
> ChickWeight[ChickWeight$Diet==4,]

3. using the function which
> ChickWeight[which((ChickWeight$Diet == 4) & (ChickWeight$Time==21)), names(ChickWeight) %in% c("weight","Time")]

Outer join recap

Given the two tables COUNTRIES and LOCATIONS, in a 1-> m relationship, with a foreign key in the LOCATIONS table.

Not all the countries have a location child record.

These two statements illustrate the use of a left outer join:

1. Display all the records of LOCATIONS, and the related records in COUNTRIES:

select c.country_name, loc.*
from locations loc
left outer join countries c 
on c.country_id = loc.country_id;

The result is equivalent (in this case) to a inner join, since no record in LOCATIONS have no parent in COUNTRIES.

2. Display all the records of COUNTRIES, and the related records in LOCATIONS (if any):

select c.country_name, loc.city
from countries c
left outer join locations loc 
on c.country_id = loc.country_id
order by c.country_name;

COUNTRY_NAME                         NVL(LOC.CITY,'NULL')          
------------------------------------ ------------------------------
Argentina                                NULL                       
Australia                                Sydney                     
Belgium                                  NULL                       
Brazil                                   Sao Paulo                  
...   

K-Nearest Neighbor Algorithm

Here's a good introduction to the k-nearest neighbor algorithm.

K-nearest neighbor is a supervised clustering algorithm. Can be used for classification and regression problems.

Example of a trivial classification problem: based on the age of the subject, determines if one like pineapple on his pizza:

The output (label) of a classification algorithm is typically represented as an integer number such as 1, -1, or 0

Example of a trivial regression problem: predict the weight of a person given their height:

Output of a regression problem is a real number

Aborting a merge

I ran into a conflict while merging, and wanted to abort my merge.

git reset --hard HEAD

excellent questionnaire en ligne à https://response.questback.com/

Bayes Theorem

In short:

P(B|A) = P(A|B) * P(A) / P(B)

see introductory articles:
https://towardsdatascience.com/naive-bayes-classifier-81d512f50a7c

R : remplace a String

I have a DataFrame sampleUvg50 with a column GNR in this format:

=TEXT(12040373;00000000)

And I would like to get rid of the parts "=TEXT(" and ";00000000)"

To obtain a string replacement in R, I can use the gsub function with a reg exp as follow:

sampleUvg50$GNR <- gsub(";\\d*\\)", "", gsub("=TEXT\\(", "", sampleUvg50$GNR));

R: select from data frame (join types):

Selecting data in Data Frame with dplyr:

Find entries in List that are in another list

df1 %>%

  inner_join(df2, by="myColumn")

Find entries in List that are not in another list

df1 %>%

  anti_join(df2, by="myColumn")

Naive Bayes Classifier

The basic asumption when using a Naive Bayes classifier is that each feature pair being classified is independant of each other and contributes equally to the outcome.

Naive Bayes classifiers have worked quite well in many real-world situations, e.g. document classification and spam filtering. 

They require a small amount of training data to estimate the necessary parameters.

Naive Bayes learners and classifiers can be extremely fast compared to more sophisticated methods.

Remove docker imagers referenced in several repositories

I recently was confronted with a problem I though needed that I delete all my docker images. So I issued the classical command:

• docker rmi $(docker images -q)

Unfortunately, it looked like some images had the same id: