Análise de sentimento do COPOM com R (parte 1: usando a base de dados)

Inspirado no artigo “Quando as palavras contam a história”, do Terraço Econômico, resolvi elaborar minha própria implementação do método, contribuindo para a comunidade. Vou mostrar então como eu fiz em alguns posts dessa série:

  1. Usando a base de dados
  2. Replicando a base de dados
  3. Analisando o sentimento

Resolvi começar pelo final, divulgando a base de dados já pronta, para que os colegas possam já começar a criar suas próprias análise, nuvem de termos, contagem de palavras, etc.

Continuar lendo

Anúncios

Curso de introdução à programação voltada aos dados

Assunto muito importante para os novos economistas e para os antigos que não querem ficar pra trás. Esse curso cobre o básico, de forma simples, focado na linguagem R, Markdown, uso de GitHub e no tratamento e visualização de dados.

Welcome to INFO-201, Technical Foundations of Informatics. This is a course at the University of Washington’s Information School, and these materials were co-developed by Michael Freeman and Joel Ross. The purpose of this course is to teach students the necessary technical skills to begin writing code to work with data. While these resources are built for students currently enrolled in the course, they have been structured to be an online resource for anyone hoping to learn to work with information using programmatic approaches.

Fonte: Technical Foundations of Informatics

Impossibilidade da escolha racional do consumidor

DOS ARQUIVOS DOS RASCUNHOS: (5 anos atrás)

Sempre me interessei pela escolha do ponto de vista do consumidor, que acho menos estudada tradicionalmente que a da empresa, e esse artigo que li na época segue um pouco essa linha misturando Simon, teoria evolucionária e Tversky.

ABSTRACT

In this paper we show that a rational consumer choice along the lines traditionally suggested might lead to paradoxical results if one considers multidimensional goods, which incorporate a series of incommensurable aspects. Thereby, we explore the similarity between the resulting paradox and Kenneth Arrow’s well known Impossibility Theorem. Based on these considerations we suggest a solution for the former problem along the lines of Herbert Simon and Amos Tversky, which might—if driven to its extreme—even provide a unique and arguably rational solution for consumer choice among multidimensional goods. Eventually, we argue that the resulting framework poses a potentially useful starting point for further developing an evolutionary theory of consumer choice.

via Journal of Evolutionary Economics, Online First™ – SpringerLink.

Valor para vantagem competitiva

DOS ARQUIVOS DE RASCUNHO: (5 anos atrás)

RESUMO. Esse link veio do escritório e não da universidade mas tem links com organização industrial pelo menos como tipologia das firmas, que pode valer a pena utilizar no lugar ou em paralelo com a do GIC-IE ou outras mais tradicionais.

Building on Thompson’s (1967) typology of long-linked, intensive, and mediating technologies, this paper explores the idea that the value chain, the value shop, and the value network are three distinct generic value configuration models required to understand and analyze firm-level value creation logic across a broad range of industries and firms. While the long-linked technology delivers value by transforming inputs into products, the intensive technology delivers value by resolving unique customer problems, and the mediating technology delivers value by enabling direct and indirect exchanges between customers. With the identification of alternative value creation technologies, value chain analysis is both sharpened and generalized into what we propose as a value configuration analysis approach to the diagnosis of competitive advantage. With the long-linked technology and the corresponding value chain configuration model as benchmark, the paper reviews the distinctive logic and develops models of the value shop and the value network in terms of primary activity categories, drivers of cost and value, and strategic positioning options.

via Configuring value for competitive advantage: on chains, shops, and networks – Stabell – 1998 – Strategic Management Journal – Wiley Online Library.

Improving the Economic Complexity Index

Interessante nova medida de complexidade. Parece um pouco difícil a replicação.

Complexity Digest

How much knowledge is there in an economy? In recent years, data on the mix of products that countries export has been used to construct measures of economic complexity that estimate the knowledge available in an economy and predict future economic growth. Here we introduce a new and simpler metric of economic complexity (ECI+) that measures the total exports of an economy corrected by how difficult it is to export each product. We use data from 1973 to 2013 to compare the ability of ECI+, the Economic Complexity Index (ECI), and Fitness complexity, to predict future economic growth using 5, 10, and 20-year panels in a pooled OLS, a random effects model, and a fixed effects model. We find that ECI+ outperforms ECI and Fitness in its ability to predict economic growth and in the consistency of its estimators across most econometric specifications. On average, one standard deviation increase in…

Ver o post original 147 mais palavras

Ciência reprodutível com R

Segue o link para uma apresentação do pessoal da Microsoft de como utilizar a linguagem R para aumentar a reprodutibilidade de sua pesquisa. Este vem sendo um tema que tenho insistido nos últimos tempos e que tem crescido bastante, com novos métodos e recomendações práticas mais claras.

Yesterday, I had the honour of presenting at The Data Science Conference in Chicago. My topic was Reproducible Data Science with R, and while the specific practices in the talk are aimed at R users, my intent was to make a general argument for doing data science within a reproducible workflow. Whatever your tools, a reproducible process:Saves time,Produces better science,Creates more trusted research,Reduces the risk of errors, andEncourages collaboration.

Fonte: Reproducible Data Science with R

Fazendo previsão com o Profeta (não, não esse)

Vejam só, o Facebook tem um time de dados e eles utilizam o R como ferramenta de estatística. Eles utilizam um pacote próprio para previsão, que utiliza técnicas bayesianas, que pega uma série com datas e dados e cria previsões só com isso. O pacote é aberto e se chama “Prophet”. Versão também disponível para Python.

Dica do blog Revolutions, que é da equipe de dados da Microsoft (!), que tem sua própria versão do R (!!) e que é disponível para nós mortais baixarem (!!!).

Facebook is a famously data-driven organization, and an important goal in any data science activity is forecasting. Now, Facebook has released Prophet, an open-source package for R and Python that implements the time-series methodology that Facebook uses in production for forecasting at scale.

Fonte: Revolutions