# Data Preparation

# General dataset properties

A dataset summarizes research data on a delimited situation in a structured manner. The research data should be prepared in such a way that easy reuse is possible. In principle, the data record should be tabular and column-based. In most cases, soil and agricultural data have a spatial reference. The spatial position of the measuring points or areas should be given in the table as detailed as possible.

A typical dataset for data transfer to the BonaRes Repository has the following properties:

Each column (attribute) of the table contains the attribute name in the first row and the attribute values in the following rows, which means the data within the table are column-oriented.
Each table or dataset must be given a short, concise name.
For widespread reuse, work should be done in English if possible.

Typically, each table contains the following standard column names:

*Table 1: Example table with typical standard column names of a data table*

# Formal criterias for datasets

The following formal criteria for tabular data especially submitted as Excel files must be met:

# Table criteria

No.	General criteria for tables
1	Tables are unformatted (do not contain graphical elements, colored lines, …)
2	There are no internal references or formulas within a table
3	Tables contain only one worksheet (avoid Excel folders with multiple sheets)

# Column criteria

No.	Criteria for column
1	Each column contains a unique attribute name (between upper and lower case not differentiated z. B. “beetles” and “Beetles”
2	Column delimiter is consistent and clearly identifies throughout the table (e. g. ”;” or “,”)
3	Column names contain no spaces (an underscore „_“ is allowed to use), special characters, umlauts, “ß” or units
4	Column names are a maximum of 30 characters long
5	Column names do not start with a number
6	Column names are only assigned once
7	Values within a column are constant (no change of reference variable or unit)

# Cell criteria

No.	Criteria for cells
1	Missing values are explicitly marked and clearly differ from “none specification "and" 0 "
2	Values in the cells contain no units, explanations or abbreviations e. g. "Under detection limit” or “< 0.00 ”
3	Only one value specified per cell
4	Cells are not connected
5	Cells do not contain separators such as “;” (decimal numbers are uniformly written with “.” or “,” and differ from column delimiter)
6	Cells contain no spaces
7	IDs are only assigned once
8	Each line is available once(there are no double lines e. g. due to copy errors)

# Preferred file formats

In case of tabular data, the prepared dataset should be submitted in the file formats:

Textfile (*.txt)
Comma separated value file (*.csv)
Excel sheet (*.xls or *.xslx).

Avoid file formats that cannot be read with common programs. For example, formats for special company software for data loggers. In addition to tables, the BonaRes Repository also publishes all formats of research data that are common in science, such as:

pictures,
videos,
texts.

The BonaRes Repository also is able to deal with complex file structures, e.g.:

MS Access
SQL
Shape files
...

In this case contact the support team of BonaRes Repository.