Creating a dataset : Step-by-step instructions

Licensed according to this deed.

Published on

Abstract

Follow these instructions to build a dataset for your research data.

 

Step What to do When to do it More details Screenshot
Step 1 Register one time You must be registered to create your own dataset. However, any user can explore the published Public Datasets  
Step 2 Log in every time you want to work on your datasets You must be logged in to create your own dataset . However, users do not have to have a login to explore the published Public Datasets. You may choose the option to stay logged in.
Step 3 Explore available datasets at least once to get familiar with the platform Go to DataCenterHub and select DATA > Discover

Users can explore published Public Datasets here. Hover pointer over column headers for column definitions.
 
 
Step 4 Create your dataset when you want to create a new dataset for your research data Go to DataCenterHub and select DATA > Create

You will see a box where you need to enter a title for your dataset. Once you provide a title for your dataset and click Create Dataset, you will see a dashboard with three main sections:

Experiment or Case Information
Experiment or Case Files
Experiment or Case Parameters


Each experiment (or case) refers to an activity (physical test, simulation, survey, etc.) in which the response of an object or specimen to defined input/excitation is documented.

Create dataset



Enter dataset title



3 main sections

 
Step 5 Add Experiment
or Case Information

to your dataset
for each new experiment or case Experiment or Case Information refers to generic information that others can use to find your dataset in the repository.

Provide a Title for each experiment. This is not necessarily the same as the title of the Dataset. It is the title of the project or group of experiments from which each experiment is obtained.

Experiment or Case ID is the ID assigned by the source(s) to the experiment.

Source refers to the name(s) of the person(s) who recorded the data.

Keywords are words describing the dataset. Users search for specific experiments in the repository with these keywords.

Latitude and Longitude are the coordinates of the location where data were recorded. These are not needed for simulations.

Compiled By refers to names of the people who compiled the data. Compiling refers to organizing the data into a dataset as opposed to recording the data during the experiment.

Compiled On is the date when the dataset was compiled (format: YYYY-MM-DD).



There are two ways to fill in, add, update, or delete information.

One way is through the Experiment or Case Information web interface. Use the Add button under the Actions column to add experiments. Once you have added an experiment you may update it by clicking on the entry you need to update. You may also delete an experiment with the Delete button under Actions.

Another way is through the Bulk Upload on the Experiment or Case Information title bar. Use this button to upload a CSV format spreadsheet in three steps:
  1. Click to get an empty CSV file with the required format. For existing datasets this file will contain existing experiment information and unique system-generated experiment IDs.
  2. Update the CSV file.
  3. Select Add & Update only or Add, Update & Delete, then select the CSV file and click Finish

Add & Update only will update experiments with matching system-generated IDs and add new experiments (without system-generated IDs).

Add, Update & Delete will update experiments with matching system-generated IDs, add new experiments (without system-generated IDs), and delete previously entered experiments not included in the CSV file being uploaded.

Experiment Title



Experiment or Case ID



Source



Keywords



Latitude and Longitude



Complied By



Compiled On



Add



Bulk upload



Get CSV file



CSV file format



Add & Update only



Add, Update & Delete



Experiment information added

 
Step 6 Upload Experiment
or Case Files

to upload or manage research data files including reports, drawings, data files with measurements, photos, videos, and audio collected during research Files of different types (reports, drawings, measurements, photos, videos, audio, etc.) are generated with each experiment. DataCenterHub has been designed to help you organize and preserve these files. They are grouped through the Experiment or Case Files interface as follows:

Report(s) are the documentation related to the experiment.

Drawings/Diagrams are the image files that are needed to interpret the data, including drawings illustrating the test set-up and the specimen or site.

Data include files (preferably in text format, e.g. *.txt, *.csv) with measurements and observations. It is recommended that data files are organized in columns with each column having a descriptive header (e.g., sensor ID, and units). Material sample tests may be stored here.

Photos, Videos, etc. are the media files including photos, videos, audio generated through research.

There are two ways to add, or update files to Experiment or Case Files.

One way is through the Experiment or Case Files web interface. Drag & Drop the file onto the Upload Files button. Another way is to click Upload Files and select the files with a file explorer.

When you want to add descriptions to the files uploaded or would like to add or delete files, click Manage to update and annotate the files. Annotations will be visible to users and will help them select files in large file groups. Investing time annotating files will help others use your data.

Upload Reports



Upload Drawings



Upload Data files



Upload Photos

 
Step 7 Add Experiment
or Case Parameters

to upload or manage the parameters used to describe the experiment The Experiment or Case Parameters interface helps you define parameters (variables) that the researcher, compiler, professional or scientific organization chose to describe the experiment in quantitative terms. Examples may include dimensions, material properties, temperature, key test results or indices, etc.

Click the drop-down menu from the Select Parameter Set to see the available pre-defined parameters. Ideally, these would be defined by professional organizations through consensus among experts. Using a pre-defined parameter set helps compare your results with other results. If you find a parameter set that corresponds to your data, select it and click on Use Parameter Set.

For flexibility, you can create your own list of parameters by clicking New Parameter Set. There are three steps to define your Parameter Set:
  1. Enter the name of your parameter set. You may add descriptions or information to your parameter set.
  2. Create a CSV file in which each column header is a parameter in your parameter set. Here is the format of the CSV file:
    row 1: column headers
    row 2: units
    row 3: descriptions
    row 4: sample data
    For row 4, you should provide at least one row of sample data to define the data types of each parameter (i.e. text, integer, decimal, URL, file, etc.). An example CSV file is provided for download.
  3. Upload your CSV file to define your Experiment or Case Parameters. Once the CSV file is uploaded, your new parameter set will be shown under User Defined when you click the drop down menu to Select Parameter Set. Select it and click Use Parameter Set to create a parameter set with your column headers.

If you do not like the parameter set you chose, click Reset and redo step 1 to 3.

Whether you used a system-defined or user-defined parameter set, you can input values for your parameter set in two ways (this works just like Experiment or Case Information – and we repeat the details here):

One way is through the web interface. Use the Add button under the Actions column to add experiment parameters. Once you have added an experiment parameter you may update it by clicking on the entry you need to update. You may also delete an experiment parameter with the Delete button under Actions.

Another way is through the Bulk Upload on the Experiment or Case Parameters title bar. Use this button to upload a CSV format spreadsheet in three steps:
 
  1. Download an empty CSV file with the required format. This file will contain the unique system-generated experiment IDs, title of the project, and experiment or case ID.
  2. Update the CSV file.
  3. Select Add & Update only or Add, Update & Delete, then select the CSV file and click Finish


Add & Update only will update experiment parameters with matching system-generated IDs.

Add, Update & Delete will update experiment parameters with matching system-generated IDs and delete previously entered experiment parameters not included in the CSV file being uploaded.

Pre-defined parameter set



New parameter set



Parameter set name



Parameter file instructions



Parameter CSV content



Upload parameter headers



Select user-defined



Use Parameter Set



Reset



Bulk upload



Get CSV for upload



Parameter data file



Adding data to file



Data file uploaded



Ready to publish!

 

Cite this work

Researchers should cite this work as follows:

  • Chungwook Sim; Santiago Pujol; Ayhan Irfanoglu; Sudheera R, Fernando; Ann Christine Catlin; Aishwarya Y Puranam; George A Howlett; Merve Usta (2015), "Creating a dataset : Step-by-step instructions," https://datacenterhub.org/resources/30.

    BibTex | EndNote

Tags