Repository for the paper submitted to Data in Brief regarding COVID19 Special Edition

Marcel Ribeiro-Dantas fdae416f28 Fix locality name varialbe 2 months ago
.dvc 3613d3cd6f Updates preprocess pipeline with several changes 3 months ago
data bb6b0d39e6 Creates pipeline for checking updates on datasets 2 months ago
documentation 3246bfecee Starts tracking with git the PDF documentation files 3 months ago
extra fdae416f28 Fix locality name varialbe 2 months ago
scripts e7599e1509 Updates datasets 24 May 2020 2 months ago
.gitignore 3246bfecee Starts tracking with git the PDF documentation files 3 months ago
DIB.Rproj 6a5dcf4680 Adds R Project file 3 months ago 6063cbbf97 Updates README with paper information 2 months ago
generate_dictionary_file.dvc b855a2711c Updates dictionary generation pipeline 3 months ago
generate_single_UN_dataset.dvc 3ebfc24496 Fix generate single UN dataset pipeline 3 months ago
preprocess.dvc 0d3f4c5748 Updates datasets (ECDC / Google Mobility) 2 months ago
update_datasets.dvc 0d3f4c5748 Updates datasets (ECDC / Google Mobility) 2 months ago

Data Pipeline

DVC Managed File
Git Managed File
Stage File
External File

How to work with this repository

Paper published in Data in Brief

Originally, this is the repository for the publication "Dataset for country profile and mobility analysis in the assessment of COVID19 pandemic" in the journal Data in Brief. You can check the paper here. Even after the publication, we still update it with the most updated data on COVID-19 and mobility. Therefore, right now this repository has more data than when the paper was published.

If you want the dataset and data dictionary just like it was on the publication, click here. If you want the most updated version of the dataset and data dictionary, click here. The source code and pipelines are also updated. If you want to see the repository just like it was on the moment of the publication, click here.

Clone the DAGsHub repository

git clone

This will copy the git repository to your computer. After that, move into the directory by typing:

cd DIB_COVID19_paper

Install DVC

If you're on GNU/Linux, you can install DVC with pip:

pip install dvc

More install options here.

If you don't want the most updated version of this repository, but the version of the publication, you should move the repository back in time by typing: git checkout tags/v1.0

Get data

If you move into data/preprocessed/ you will see that the folder is empty. Yes, it is, dataset files should not be there, afterall they are not tracked by git. The same thing applied to data/raw. It is not recommended to track big files and objects with git. The documentation folder folder will not be there either (though you can see the documentation.dvc file there, which means DVC tracks it ).

dvc pull

It's the same reasoning as if you wanted to get the latest tracked code by git (git pull). Now you will find the files in the data/raw and data/preprocessed files, and also the documentation files in the documentation folder.

Reproduce the pipeline

Let's say you have an updated version of one or more of the raw files, on the same format (column names did not change, for example), and you want to reproduce the pipeline (generate an updated version of the preprocessed file). From the root directory of the repository, run:

dvc repro preprocess.dvc

DVC will automatically notice that one of the raw files changed and therefore it will reproduce the pipeline and generate a new output file in the data/preprocessed folder. If nothing changed, dvc will realize there is no reason to reproduce the pipeline and you will see the following message:

Data and pipelines are up to date.

Well, maybe you want to force a reproduction fo the pipeline because... well, you want to see how it would be like. For this, you should run:

dvc repro -f preprocess.dvc