The following data were selected from the set "F" of the public database of the North Sea, released in NL Oil and Gas Portal a website sponsored by the Dutch Government providing information on oil, gas and geothermal energy exploration and production in the Dutch sector of the North Sea continental shelf.

The original data in NL Oil and Gas Portal are text files (.LAS) containing the depth and various well logs for each well and an image file (.TIFF) containing the lithology information. The lithology information as images is an obstacle to use it directly for computational tools.

Some files were preprocessed to facilitate their use. They are available in the following format:

Tab-separated values

ASCII text file containing one record per line. All records are following the same order of features separated by a TAB, and the first line is a header with the features' names.

WELL    DEPT    GR      DT      RHOB    DRHO    NPHI    LITO
F01-01  1594    20.28   120.74  2.23    -0.01   0.32    mudstone
F13-01  1788.2  11.74   51.18   2.95    -0.02   -0.01   anhydrite
F16-05  4549.9  81.77   58.08   2.63    0.23    0.11    sandstone

Eleven wells were selected for reformatting. These wells belong to the set "F" which has 108 wells. The eleven wells were chosen based on their positions and occurrence of the same well logs. The following image represents the location of the 108 wells, and the selected wells are in red.

The files contain information of seven features:

  • WELL: number of the well

  • DEPT: depth value to well logs

  • GR: gamma-ray value which measures the total radioactivity of the geological formation

  • DT: sonic value, representing the difference in transit times of a mechanical wave through the rocks

  • RHOB: bulk density value, it is the density of the rock, including fluid-filled pore space

  • DRHO: delta rho value, representing the presence of mud and holes that cause deviations in the density curve

  • NPHI: neutron porosity value, measuring the initial estimate of porosity

  • LITO: rock type recovered in the image file

The information about features were obtained from "Fundamentos de Engenharia de Petróleo" (Thomas, 2004).


All the files contain samples from wells: F01-01, F03-03, F03-07, F06-03, F08-01, F11-02, F12-01, F13-01, F16-05, F17-03 e F18-11. The second file is a subset of the first.


These data were collected and formatted in the scope of the master’s dissertation "Uma análise da classificação de litologias usando SVM, MLP e métodos Ensemble". In this work, the files were also converted to ARFF format to use in Weka. ARFF files can be downloaded here (.zip).