Comparto aquí el material utilizado en la sesión titulada “Análisis de correspondencias con R: aplicación a datos de encuestas” que impartí recientemente en la Facultad de Ciencias de la Salud de la Universidad Rey Juan Carlos. Al seminario asistieron alumnxs de distintos grados y alguno de posgrado, sobre todo de psicología, por lo que me pareció adecuado enfocarlo a datos de encuestas. Es interesante ver cómo tratar datos “reales” que tenemos a disposición en forma de microdatos, sobre todo porque en las clases de grado trabajamos con ejemplos muy específicos, a menudo artificiales.
En cuanto a la técnica, el análisis de correspondencias, es aplicable a cualquier disciplina, desde las ciencias sociales hasta la biotecnología o la bioinformática. En este caso en concreto, en el ámbito de la bioestadística se analizan datos de la encuesta europea de salud, pero se podría aplicar a datos recogidos en hospitales, en granjas, cultivos, etc., y realizar interesantes análisis para la investigación en estos ámbitos.
Como no podía ser de otra forma viniendo de mí, el material está realizado con el software estadístico R, y se encuentra en un repositorio github con instrucciones para su descarga y uso inmediato desde RStudio:
Today a user of my R package SixSigma
and reader of its motivating book, Six Sigma with R, made my day. The book was published in 2012 by Springer, and it was my first scientific publication (even before any article). Later (2015) Quality Control with R came with new and improved content.
Back to what I wanted to talk about: Paul, a user from a company in Germany, reported a “bug” in a function of the package. After telling him how I was going to solve it and some new plans to improve the whole thing, he gave me two senteces that encourage me to continue working on the topic:
I use your package on daily basis, as user I really appreciate your work
Regarding the book, I think it should be present in each personal library of a black belt
So, someone working in the “real world” uses on daily basis the open source software that one has developed, and says that your book should be a must for a black belt (see wikipedia for newcomers to Six Sigma). Sometimes one does things apparently nonsense, but later the reason appears.
Yeah, all right. It’s free software. I do not earn anything, beyond a few lines in the curriculum to fight in the academic career, and the satisfaction of doing things that serve society. Maybe I could be making more money. But what I would really like is to get funding to hire young researchers and take this development further and faster, so that we can bring these advances to companies nearby.
Part of the conference series Breakthrough science and technologies. Transforming our future.
23 January 2018, The Royal Society, London, UK
Notes from my own perspective
This was an extraordinary event for those working or researching on the energy storage topic. In just one day, a mix of science, technology and policy talks showcased the current state of the art and forthcoming challenges on energy storage. It was really enlightning for my current developments in one of my research lines.
All the speakers gave nice presentations on their topics. I could catch the following ideas and concepts to further explore in my research.
From the introduction by Professor Peter Bruce:
Large scale storage is a cornerstone (heard several times during the conference)
It’s not windy all the time (of course not)
Small scale storage is also coming to the play
From the energy policies presentation by Joan MacNaughton:
The 3 Ds: Decarbonisation, Disitisation, Decentralisation
Scenarios projections IAE: NPS (New Policy Scenario) and SDS (Sustainable Developement Scenario
Policy an regulation is focusing on outcomes, such as decarbonisation, reliability, cost optimisation, and protection of consumers and their data.
Such policies will be underpinned by storage.
From Jorge Pikunic talk:
The key concept of flexibility
Optimisation of different technologies
Energy consumers taking the control
Availability of new technologies
Challenge with EVs: network stability. Other types of batteries may help, through, again, flexibility, leading to sustainability and resilience.
From Faraday’s challenge talk:
Lifetime and performance
From Alexander Slocum (impressive) talk:
100$/kW seems like a milestone (datum in several talks)
From Professor Ian Metcalfe talk:
Chemical enery is still there
Hydrogen storage, from current combustion
Adiabadic temperature (c.f. loss recovery in my work)
From Dr. Christos Markides talk:
Energy is not only electricity
Thermal energy is not only hot
Thermal storage can be an intermediate for electricity
From Dr. Jeffrey Chamberlain talk:
Suite of (storage) technology
I did not take notes from Professor Laura Diaz, Dr. Ben Irons and Chris Brown, but they were also quite interesting for my research.
Further discussions during the breaks and reception were also absolutelly enriching. Sara de la Serna (sse Enterprise) talk to me about Vehicle to Grid (V2G) which I knew nothing about, and sounds a very attractive topic. Fidel Tamayo works at eTran on mobility, and their project in the taxi sector looks very interesting. Discussions with Koichi GOTO, from JAPAN NUS CO., LTD, were really nice. Not only we talked about energy, but also about analytics on the Internet of Cows (and pigs, and so on). Finally, Dr. Timur Yunosov (University of Reading) and I shared the importance of building generic models for energy, even though it is difficult (but not impossible) to have them work accurately in specific technologies.
Disclaimer: this post is more a way to record my experience in the event for future reference than an attempt to summarise the whole thing. If you reader find any improvements to this, or something I should change, please comment the post.
I have made the changes after the feedback received by package users from industry (thank you very much Luc and Austin). It is really encouraging such input and motivates me to continue developing the package. In fact, my research group is looking for funding to develop a complete R infrastructure for quality control and improvement standardised statistical methods. There is a real need for that, see for example one of the messages from a real world user:
This morning I discussed Gage R&R with my “Minitab” collaegue. As we now calculate 144 Gage R&R’s, he could not even generate those in Minitab, the advantages of R are becoming more and more clear to him. […]
Friends at the International Institute for Applied Systems Analysis (IIASA) have kindly published a news item featuring my recent work. I spent there a pre-doc stay during the summer of 2013, allowing me to get e European Doctorate. I had the opportunity to interact with the great Yuri Ermoliev. It was in the framework of the competitive Young Scientists Summer Program (YSSP) with researchers from all over the World.