396
Box 15.1 An Example of Folder Structure
It can be advantageous for research groups and institutions to implement common data standards and a file structure that forms the backbone of a data
organization scheme and does not have to be discussed for every new project.
Data standards also promote reproducible research. One option is to set up file
directories separated by content. The example below structures a higher-level
directory (e.g., the project directory) into docu_work, docu_pub, orig_data,
data_work, data_pub, gis_work, gis_pub, maps, and printout folders. The key
feature of this structure is the distinction between work folders, containing
work in progress and intermediate results, and pub folders, containing the
final versions and results. After the completion of a project, the orig_data,
maps, printout, and all pub folders are archived (publicly and internally),
while all work folders get deleted from a public or shared drive. Backups are
kept at a certain frequency for a certain amount of time (e.g., daily backups
for a year, biweekly backups for 3 years, monthly backups for 10 years), and
personal copies can of course be kept as long as necessary. A folder structure
like this makes archiving data easy because at each project stage it is clear
which data sets, documents, and products will be preserved. The pub, orig_
data, and maps folders should contain everything a person without knowledge
about the project would need to repeat the analysis, including basic project
background and workflow descriptions in the docu_pub folder, but nothing
unnecessary, such as intermediate results. One testable reproducibility goal
could be that a person with appropriate analytical skills but without any information about the project would be able to re-create and explain a main result,
including the rationale behind the analysis, after 1 workday without any external help. For GIS heavy projects, it may make sense to separate analysis,
results, and products based on geospatial data from those using other data
sources. In this example, map products are reproducible based on data from
gis_pub and the layouts found in the maps folder.
Short Description of Contents and Management
of Example Folder Structure
Folder
type
Content description and management
docu
folders
These are for the proposal, project descriptions, documentation of workflows,
planning documents, minutes of meetings, manuscripts, photos, etc. File names
could, for example, start with the ISO date followed by a descriptive name. Files
are usually organized into subfolders. Typical file formats include .doc, .txt.,
.pdf, and .tiff. The docu_work folder is deleted after the project is finished. The
docu_pub folder contains the final versions in a non-editable format, such as
.pdf. It should contain the essentials of the project background and all workflows
needed to repeat the analysis; detailed descriptions should either be left out or
clearly flagged, e.g., as “additional information.” The docu_pub folder also
includes information about the use of corporate or project identity styles (use of
logos, colors, fonts, etc.). This folder gets archived after the project is finished
A. K. Schweiger
Box 15.1 An Example of Folder Structure
It can be advantageous for research groups and institutions to implement common data standards and a file structure that forms the backbone of a data
organization scheme and does not have to be discussed for every new project.
Data standards also promote reproducible research. One option is to set up file
directories separated by content. The example below structures a higher-level
directory (e.g., the project directory) into docu_work, docu_pub, orig_data,
data_work, data_pub, gis_work, gis_pub, maps, and printout folders. The key
feature of this structure is the distinction between work folders, containing
work in progress and intermediate results, and pub folders, containing the
final versions and results. After the completion of a project, the orig_data,
maps, printout, and all pub folders are archived (publicly and internally),
while all work folders get deleted from a public or shared drive. Backups are
kept at a certain frequency for a certain amount of time (e.g., daily backups
for a year, biweekly backups for 3 years, monthly backups for 10 years), and
personal copies can of course be kept as long as necessary. A folder structure
like this makes archiving data easy because at each project stage it is clear
which data sets, documents, and products will be preserved. The pub, orig_
data, and maps folders should contain everything a person without knowledge
about the project would need to repeat the analysis, including basic project
background and workflow descriptions in the docu_pub folder, but nothing
unnecessary, such as intermediate results. One testable reproducibility goal
could be that a person with appropriate analytical skills but without any information about the project would be able to re-create and explain a main result,
including the rationale behind the analysis, after 1 workday without any external help. For GIS heavy projects, it may make sense to separate analysis,
results, and products based on geospatial data from those using other data
sources. In this example, map products are reproducible based on data from
gis_pub and the layouts found in the maps folder.
Short Description of Contents and Management
of Example Folder Structure
Folder
type
Content description and management
docu
folders
These are for the proposal, project descriptions, documentation of workflows,
planning documents, minutes of meetings, manuscripts, photos, etc. File names
could, for example, start with the ISO date followed by a descriptive name. Files
are usually organized into subfolders. Typical file formats include .doc, .txt.,
.pdf, and .tiff. The docu_work folder is deleted after the project is finished. The
docu_pub folder contains the final versions in a non-editable format, such as
.pdf. It should contain the essentials of the project background and all workflows
needed to repeat the analysis; detailed descriptions should either be left out or
clearly flagged, e.g., as “additional information.” The docu_pub folder also
includes information about the use of corporate or project identity styles (use of
logos, colors, fonts, etc.). This folder gets archived after the project is finished
A. K. Schweiger