Important: Open-FF data is based on FracFocus data. FracFocus has outlined their “Terms of Use” on their download page. Those requirements are relatively generous, but users of Open-FF data should be familiar with them.
You can use Open-FF’s Data Browser to explore the data set from many perspectives without downloading data. We suspect that many users’ questions can be answered with this tool. However, if you wish to work directly with the data set for more in-depth or complex analyses, you may download the Open-FF data.
Flat-file download
A full data set is available by simple download from here. The file is in ‘parquet’ format and is over 500Mbytes. See the Data Dictionary for a description of fields. A full ‘CSV’ version would be much larger download due to the format. Until we can develop a different online solution for other flat-file formats, please contact us and we’ll do what we can to get you what you need.
Database generation
Alternatively, if you wish to create a database version (instead of the flat file version), you can use the following parquet tables to construct a database:
| Link to parquet file | Description |
|---|---|
| bgCAS | (key: bgCAS); Basic info about the specific chemical |
| cas_ing | (keys: CASNumber and IngredientName); Open-FF’s translation table of FracFocus’s chemical ID and name. |
| chemrecs | (key: reckey) Every record in FracFocus; chemical ID, quantity, TradeName, Supplier, etc. |
| disclosure | (key: DisclosureId) Dates, location, Operator, Water Volume, TVD, etc |
| water_source | (key: DisclosureId) Recent addition to FF: six categories of where the water comes from. |
Examining the Open-FF data dictionary should help. Below each field name in the dictionary is the table(s) the field is in.
Important: An additional field is needed for proper filtering. In the flat file, it is named “in_std_filtered” and it is used primarily to remove disclosures and records that are flagged by Open-FF as duplicates. You can create that field yourself using “dup_rec” and “is_duplicate” (see the dictionary). Roughly:
in_std_filtered == (NOT dup_rec) AND (NOT is_duplicate)
