Personal thoughts mainly about AI and Management. You can also select a specific tag to read all post related to it.
I’m a bit tired of writing the usual gushing posts about how amazing and empowering it was to be part of event X, Y, or Z. Instead, here’s a short, candid note on what I actually took home after sharing my pitch — “Bringing Internal Audit into the AI Age” — with a large, well‑qualified crowd of auditors at the IIA Annual Conference in Luxembourg.
Tre anni fa, circa, sono stato nominato per la prima volta Head of di un team, che oggi è composto di circa 20 persone. Ho sempre avuto una certa passione per il management, nel senso che nel tempo sono andato ragionando su quali fossero i modi migliori per far lavorare bene insieme le persone. Se lo sto facendo bene o male chiedetelo ai fantastici #ADAAPeople, nel frattempo queste sono tre cose che ho imparato e che condivido per capire che ne pensate e perché possano essere utili a chi sta iniziando quel fantastico (e faticoso e infinito) viaggio che è diventare manager.
For personal reasons I am trying to track the number of NCOV-19 confirmed cases in Italy as well as the number of deaths (since I live in Italy, it is not difficult to guess the personal reason…). I am thus regularly monitoring news from the italian official sources like “Regione Lombardia” and “Protezione Civile”.
PaletteR, the package that allows you to create an optimized palette from an image, has been staying around for nearly two years, and #rstats users have made a lot of great stuff with it. I had therefore took the time to collect what I have found around the web, just to celebrate this beauty. You can find them below in a slideshow.
I recently came across the good talk by the always good Yiuhi Xie at the Rstudio conference about Pagedown. You can see it by yourself reaching out the rstudio website or clicking the image below:
Pagedown is a newly released package ( still in experimentary status) with a really promising mission: help you build beautiful PDFs documents from our beloved Rmarkdown. More precisely it takes Rmarkdown files and render them into html files already “paged” and ready to be saved/converted into PDF.
One of the most frustrating issues I face in my professional life is the plentitude of ineffective reports generated within my company. Wherever I look around me is plenty of junk charts, like barplot showing useless 3D effects or ambiguous and crowded pie charts. I do understand the root causes of this desperate state of the art: people have always less time to dedicate to reports crafting, and even less to dedicate to their plot. In the crazy and speedy-going working life, my colleagues have no time and for learning data visualization principles. Even so, this remains quite a big problem since a lot of time and money-wasting consequences come from poorly crafted reports and plots:
The greatest part of my time at working is spent looking at those great piece of statistical machinery that credit risk models are. That is why I was recently required to prepare a short course about the new definition of default, which is goign to be applied from 01/01/2021 on.
It is quite a unexplored topic in the sector and I haven’t found a lot of teaching material online, that is why I tought to share here the brief deck of slide I produced for the occasion. Enjoy and feel free to comment.
I live in Italy, and more precisely in Milan, a city known for fashion and design events. During a lunch break I was visiting the Pinacoteca di Brera, a 200 centuries old museum. This museum is full of incredible paintings from the Renaissance period. During my visit I was particularly impressed from one of them: “La Vergine con il Bambino, angeli e Santi”, by Piero della Francesca.
I personally really appreciate the InstallR package from Tal galilli, since it lets you install a great number of tools needed for working with R just running a function.
I have recently completed a great reading: Edward Tufte’s The visual display of quantitative information. In the dataviz realm, this is some kind of fundamental book. This book was some kind of structural break in the history of data visualization. In ’70s and ’80s graphics were considered a way to entertain less educated readers. Their ability to make available new insights and communicate them effectively was underestimated.
It has being nearly half an year since the last post about workflower was out, why did I stay so silent for that long?
I have three major updates to explain the silence:
📚 Good guys at Packt publishing asked me to write one more book about R and data mining, I suppose this is because the first one was well received
📦 I spend my spare time working on updateR so to get it ready to go on CRAN. We make it so shining and bright that it got noticed by our beloved Tal Galili and we are now working togheter to merge it into his great package installR.
we all know R is the first choice for statistical analysis and data visualisation, but what about big data munging? tidyverse (or we’d better say hadleyverse 😏) has been doing a lot in this field, nevertheless it is often the case this kind of activities being handled from some other coding language. Moreover, sometimes you get as an input pieces of analyses performed with other kind of languages or, what is worst, piece of databases packed in proprietary format (like .dta .xpt and other). So let’s assume you are an R enthusiast like I am, and you do with R all of your work, reporting included, wouldn’t be great to have some nitty gritty way to merge together all these languages in a streamlined workflow?
this short post is exactly what it seems: a showcase of all ggplot2 themes available within the ggplot2 package. I was doing such a list for myself ( you know that feeling …“how would it look like with this theme? let’s try this one…”) and at the end I thought it could have be useful for my readers. At least this post will save you the time of trying all differents themes just to have a sense of how they look like.
I am really enjoying Uefa Euro 2016 Footbal Competition, even because our national team has done pretty well so far. That’s why after browsing for a while statistics section of official EURO 2016 website I decided to do some analysis on the data they share ( as at the 21th of June).
Just to be clear from the beginning: we are not talking of anything too rigourus, but just about some interesting questions with related answers gathered mainly through data visualisation.
I was crafting this checklist for my personal use, and then I found myself thinking: why should’nt I share this useful handful of bullets with my readers? So here we are, find below an useful checklist for your weekly review. The checklist is derived directly from the official GTD book by our great friend David Allen. The greatest quality of the checklist is the minimalist approach: just what you really need to read is written within each point, so that you get through your review as quick as possible. Enjoy!
Ah, writing a blog post! This is a pleasure I was forgetting, and you can guess it looking at last post date of publication: it was around january... you may be wondering: what have you done along this long time? Well, quite a lot indeed:
changed my job ( I am now working @ Intesa Sanpaolo Banking Group on Basel III statistical models)
became dad for the third time (and if you are guessing, it’s a boy!)
The WordPress.com stats helper monkeys prepared a 2015 annual report for this blog.
Here’s an excerpt:
The concert hall at the Sydney Opera House holds 2,700 people. This blog was viewed about **8,800** times in 2015. If it were a concert at Sydney Opera House, it would take about 3 sold-out performances for that many people to see it.
This is not actually a real post but rather a code snippet surrounded by text.
Nevertheless I think it is a quite useful one: have you ever found yourself writing a function where a data frame is created, wanting to name that data frame based on a custom argument passed to the function?
For instance, the output of your function is a really nice data frame name in a really trivial way, like “result”.
It was around midnight here in Italy: I shared the code on Github, published a post on G+, Linkedin and Twitter and then went to bed.
In the next hours things got growing by themselves, with pleasant results like the following:
https://twitter.com/DoodlingData/status/635057258888605696
The R community found ramazon a really helpful package.
And I actually think it is: Amazon AWS is nowadays one of the most common tools for online web applications and websites hosting.
Because Afraus received a good interest, last month I override shinyapps.io free plan limits.
That got me move my Shiny App on an Amazon AWS instance.
Well, it was not so straight forward: even if there is plenty of tutorials around the web, every one seems to miss a part: upgrading R version, removing shiny-server examples… And even having all info it is still quite a long, error-prone process.
All this pain is removed by ramazon, an R package that I developed to take care of everything is needed to deploy a shiny app on an AWS instance. An early disclaimer for Windows users: only Apple OS X is supported at the moment.
The last Report to the Nation published by ACFE, stated that on average, fraud accounts for nearly the 5% of companies revenues.
on average, fraud accounts for nearly the 5% of companies revenues
Projecting this number for the whole world GDP, it results that the “fraud-country” produces something like a GDP 3 times greater than the Canadian GDP.
As I am currently working on a Fraud Analytics Web Application based on Shiny (currently on beta version, more later on this blog) I found myself asking: wouldn’t be great to add live chat support to my Web Application visitors?
It would indeed!
[caption id=“attachment_490” align=“aligncenter” width=“200”]
an ancient example of chatting - Camera degli Sposi, Andrea Mantegna 1465 -1474[/caption]
I know, we are not talking about analytics and no, this is not going to set me as a great data scientist… By the way: have you ever wondered how to list all files and folders within a root folder just hitting a button**?**
I have been looking for something like that quite a lot of times, for instance when asked to write down an index of all the working papers pertaining to a specific audit ( yes, **I am an auditor, **sorry about that): really time-consuming and not really value-adding activity.
Around 1938 Frank Benford, a physicist at the General Electrics research laboratories, observed that logarithmic tables were more worn within first pages: was this casual or due to an actual prevalence of numbers near 1 as first digits?
Pushing to my Github repository directly from the Rstudio project, avoiding that annoying “copy & paste” job. Since it is one of Best Practices for Scientific Computing, I have been struggling for a while with this problem. Now that I managed to solve the problem, I think you may find useful the detailed tutorial that follows. I am not going to explain you the reason why you should use Github with your Rstudio project, but if you are asking this to yourself, you may find useful a **Stack Overflow discussion **on the topic.
After all, I am still an Internal Auditor. Therefore I often face one of the typical internal auditors problems: understand links between people and companies, in order to discover the existence of hidden communities that could expose the company to unknown risks.
In order to address this problem I am developing Linker, a lean shiny app that take 1 to 1 links as an input and gives as output a network map:
If you have a blog you may want to discover how your website is performing for given keywords on Google Search Engine. As we all know, this topic is not a trivial one.
Problem is that the analogycal solution would be quite time-consuming, requiring you to search your website for every single keyword, on many many pages.
Feeling this way?
[caption id=“attachment_273” align=“aligncenter” width=“300”]
“Pain and fear, pain and fear for me” - Oliver Twist[/caption]
I reproduce here below principles from the amazing paper Best Practices for Scientific Computing, published on 2012 by a group of US and UK professors. The main purpose of the paper is to “teach” good programming habits shared from professional developers to people that weren’t born developer, and became developers just for professional purposes.
Scientists spend an increasing amount of time building and using software. However, most scientists are never taught how to do this efficiently
1. _a program should not require its readers to hold more than a handful of facts in memory at once_
2. _names should be consistent, distinctive and meaningful_
3. _code style and formatting should be consistent_
4. _all aspects of software development should be broken down into tasks roughly an hour long<!-- more -->_
1. _rely on the computer to repeat tasks_
2. _save recent commands in a file for re-use_
3. _use a build tool to automate scientific workflows_
1. _software tools should be used to track computational work automatically_
1. _work in small steps with frequent feedback and course correction_
1. _use a version control system_
2. _everything that has been created manually should be put in version control_
1. _every piece of data must have a single authoritative representation in the system_
2. _code should be modularized rather than copied and pasted_
3. _re-use code instead of rewriting it_
1. _add assertions to programs to check their operation_
2. _use an off-the-shelf unit testing library_
3. _use all available oracles when testing programs_
4. _turn bugs into test cases_
5. _use a symbolic debugger_
1. _use a profiler to identify bottlenecks_
2. _write code in the highest-level language possible_
1. _document interfaces and reasons, not implementations_
2. _refactor code instead of explaining how it works_
3. _embed the documentation for a piece of software in that software_
1. _use pre-merge code reviews_
2. _use pair programming when bringing someone new up to speed and when tackling particularly tricky problems_
if you want to discover more, you can download your copy of Best Practice Scientific Computing here below
This simple tutorial will show you how to download data into an excel spreadsheet, creating a web query.
![Image [9]](https://andreacirilloblog.files.wordpress.com/2014/10/image-9.png?w=300)
as part of the** excel functions in R,** I have developed this custom function, reproducing the excel right() function in th R language. Feel free to copy and use it.
[code language=“r”] right = function (string, char){ substr(string,nchar(string)-(char-1),nchar(string))} [/code]
you can find other function in the Excel functions in R post.
as part of the excel functions in R, I have developed this custom function, emulating the excel left() function in th R language. Feel free to copy and use it.
you can find other function in the** Excel functions in R post**.
I have started my “data-journey” from Excel, getting excited by formulas like VLookup(), right() and left().
then datasets got bigger, and I discovered that little spreadsheets were not enough, and look for something bigger and stronger, eventually coming to R.
But as you know, ones never forget the first love.
So, for fun and for practice, I have written down some of excel functions in R.
I hope you will enjoy.
Great tutorial on text mining with twitter byPaeng Angnakoon
[youtube=http://youtu.be/mJVcANlkxU8]
Following the post about %in% operator, I received this tweet: https://twitter.com/benwhite21/status/510520550553165824
I gave a look to the code kindly provided by Ben and then I asked myself: I know dplyr is a really nice package, but which snippet is faster?
to answer the question I’ve put the two snippets in two functions:
#Ben snippet dplyr_snippet =function(object,column,vector){ filter(object,object[,column] %in% vector) } #AC snippet Rbase_snippet =function(object,column,vector){ object[object[,column] %in% vector,] }
Then, thanks to the great package microbenchmark, I made a comparison between those two functions, testing the time of execution of both, for 100.000 times.
Some time ago I was looking for an easy way to put some math writing within my Evernote notes trough my Mac device. Even if there is no official solution to the problem and the feature request is still pending within Evernote dedicated forum, I finally came out with a very simple way to solve your problem out.
you haveto subset a data frame using as criteria the exact match of a vector content.
for instance:
Example: sales records, each record is a deal.
The vector is a list of selected customers you are interested in.
I have just saturated all my PC:
full is the 4gb RAM
and so is the CPU (I7 4770 @3.4 GHZ)
which is my secret?
the doParallel package for R on mac
The package lets you make some very useful parallel computation, giving you the possibility to use all the potentiality of your CPU.
As a matter of fact, the standard R option is to use just on of the cores you have got on your PC.
I’ve been recently asked to analyze some Board entertainment expenditures in order to acquire sufficient assurance about their nature and responsible.
In response to that request I have developed a little Shiny app with an interesting reactive Bubble chart.
The plot, made using ggplot2 package, is composed by: a categorical x value, represented by the clusters identified in the expenditures population A numerical y value, representing the total amount expended Points defined by the total amount of expenditure in the given cluster for each company subject. Morover, point size is given by the ratio between amount regularly passed through Account Receivable Process and total amount of expenditure for that subject in that cluster.
I live in Italy, and more precisely in Milan, a city known for fashion and design events. During a lunch break I was visiting the Pinacoteca di Brera, a 200 centuries old museum. This museum is full of incredible paintings from the Renaissance period. During my visit I was particularly impressed from one of them: "La Vergine con il Bambino, angeli e Santi", by Piero della Francesca.
If you see this painting you will find a profound of colours with a great equilibrium between different hues, the hardy usage of complementary colours and the ability expressed in the "chiaroscuro" technique. While I was looking at the painting I started, wondering how we moved from this wisdom to the ugly charts you can easily find within today's corporate reports ( find a great sample on the WTF visualization website)
This is where Paletter comes from: bring the Renaissance wisdom and beauty within the plots we produce every day.
PaletteR is a lean R package which lets you draw from any custom image an optimized palette of colours. The package extracts a custom number of representative colours from the image. Let's try to apply it on the "Vergine con il Bambino, angeli e Santi" before looking into its functional specification.
library(devtools)
install_github("andreacirilloac/paletter")
_image_path_ arg
- specify tcolorser of colours we want to draw specifying the _number_of_colours_ attribute
- make clear if we need a palette for quantitative or qualitative variables, using the _type_of_variable_ arg.create_palette(image_path = "~/Desktop/410px-Piero_della_Francesca_046.jpg",
number_of_colors =20,
type_of_variable = “categorical")
and here it is the output:
As you see the palette drawn contains all the most representative colours, like the red of the carpets or the wonderful blue of San Giovanni Battista on the left of the painting.
kmeans algo on it and draw a sample of representative colours
- move to HSV colour space
- remove too bright and too dark colours leveraging HSV colour system properties
- further sample colours to select the most "distant" ones.painting <- readJPEG(image_path)this will generate an array having for each point within the image both the cartesian coordinates and the R, G and B values of the related colours. We now apply some statistical learning on the array, to select most representative colours and create an optimized palette.
What was wrong with the palette employed? We can pick at least three answers:
rgb2hsv() to the values of R, G and B.
brightness_stats <- boxplot.stats(sorted_raw_palette$v) saturation_stats <- boxplot.stats(sorted_raw_palette$s)Once being done with that we remove the lowest and highest of both. This lets us fix the first two problems observed in the first palette: too bright and too dark colours. What about the third problem?
You would definitely say the first and the second are more distant from each other than the second and the third. You would definitely be right, but how to make our PaletteR as cleaver as you?
This is simply done within the HSV space leveraging the Hue attribute. As we have seen HSV hues are placed along a circle in a visually reasonable way. This means that a hue of 40 (which is some kind of orange) is way more distant from a hue of 100 (green) than a hue of 90 is ( another green).
Knowing this we just have to select from the first set of colours coming from kmeans a second subset of colours selected as the most _distant_. This will let us avoid employing colours appearing too similar.
How to do this? The current version of paletteR does it:
Isn't that better than the previous one?
_create_palette_ function is a vector of hex codes (another way of codify colours, more on the Wikipedia page).
You therefore have to pass it to your ggplot plot employing scale_color_manual().A small side note: be sure to select a number of colours equal to the number of variables to plot.
Let's apply our palette by Raffaello with an hyphotetical plot:
colours_vector <- create_palette(image_path = image_path,
number_of_colors =32,
type_of_variable = “categorical")
ggplot(data = mtcars, aes(x = rownames(mtcars),y = hp,color = rownames(mtcars),
fill = rownames(mtcars))) +
geom_bar(stat = 'identity') +
scale_color_manual(values = colours_vector) +
scale_fill_manual(values=colours_vector)+
theme_minimal() +
guides(size = FALSE) +
theme(legend.position = "bottom") +
labs(title = "disp vs hp")+
coord_flip()
Which will produce:
paletter is quite a young package, nevertheless it already catched some interest (I was also invited to give a speach about it, you can watch it online).
This is because of:
PaletteR has been staying around for nearly two years, and #rstats user have made a lot of great stuff with it. I therefore took the time to collect what I have found around the web. You can find them below in a slideshow.