PresentationPDF Available

How to write a scientific paper A data-scientist approach

Authors:
  • Critical Centrality Institute

Abstract

A brief guidance on how to write a scientific paper from a data scientist point of view.
How to write a scientific paper
A data-scientist approach
Juan C. Correa, PhD
Fundaci´
on Universitaria Konrad Lorenz, Bogot´
a - Colombia
kjuanc.correan@konradlorenz.edu.co
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 1 / 8
What is a paper? The data-scientist approach
Readers will find a clear indication of where to find the raw data
and the computational scripts (e.g., GitHub, Protocols.io).
Graphics will be weird for non-data-scientists. Data scientists will
know the specific libraries that were used to generate them (e.g.,
ggplot2, seaborn, matplotlib).
Statistics are well-explained along with its graphics (e.g., statistical
significance does not imply large effect).
Abstracts have two versions: the prose-based version and the
graphic-based version.
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 2 / 8
What is a paper? The data-scientist approach
Readers will find a clear indication of where to find the raw data
and the computational scripts (e.g., GitHub, Protocols.io).
Graphics will be weird for non-data-scientists. Data scientists will
know the specific libraries that were used to generate them (e.g.,
ggplot2, seaborn, matplotlib).
Statistics are well-explained along with its graphics (e.g., statistical
significance does not imply large effect).
Abstracts have two versions: the prose-based version and the
graphic-based version.
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 2 / 8
What is a paper? The data-scientist approach
Readers will find a clear indication of where to find the raw data
and the computational scripts (e.g., GitHub, Protocols.io).
Graphics will be weird for non-data-scientists. Data scientists will
know the specific libraries that were used to generate them (e.g.,
ggplot2, seaborn, matplotlib).
Statistics are well-explained along with its graphics (e.g., statistical
significance does not imply large effect).
Abstracts have two versions: the prose-based version and the
graphic-based version.
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 2 / 8
What is a paper? The data-scientist approach
Readers will find a clear indication of where to find the raw data
and the computational scripts (e.g., GitHub, Protocols.io).
Graphics will be weird for non-data-scientists. Data scientists will
know the specific libraries that were used to generate them (e.g.,
ggplot2, seaborn, matplotlib).
Statistics are well-explained along with its graphics (e.g., statistical
significance does not imply large effect).
Abstracts have two versions: the prose-based version and the
graphic-based version.
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 2 / 8
Graphical Abstract
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 3 / 8
Which are the recommended tools?
Regardless the preferred tool for producing the paper,
the compilation of its parts (i.e., results, graphics, and references) is
automated in R with latex documents via Sweave or with PyLatex in
Python.
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 4 / 8
How to write it?: The Intro
The introduction should be concise. It should follow Clearbout’s
formula, consisting of providing a review, the claim, and the agenda.
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 5 / 8
How to write it? The Method
The method should provide the specifics about data collection
procedure (institutional repositories, API’s, etc). It should provide the
date when the data was obtained and if it required some sort of
preprocessing.
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 6 / 8
How to write it? The results
Here, the author shows off all of his talents related to data
visualization. Don’t hesitate to mention readers where they can find
your codes. They will learn how to generate your elegant graphics with
your data.
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 7 / 8
How to write it? The discussion
Rephrase the aim of your paper in the first sentence. Then, number
the contributions of your work compared with previous studies and
elaborate upon each of these contributions. Finally, suggest possible
relationships between your work (data) and others’ works (data).
Juan C. Correa (Twitter: @jcorrean) Writing a paper: A data-scientist approach Bogot´
a, March, 2018 8 / 8
ResearchGate has not been able to resolve any citations for this publication.
ResearchGate has not been able to resolve any references for this publication.