COT is the acronym of Commitments of Traders. COT data are published reports to help the public understand market dynamics of the main financial instruments. Learn more here.
jCOT is a Python script that allows you to bulk download all of these data from 1986 to the present. It also allows you to update the current year's data by integrating them when new data are released. And finally it allows you to perform simple search queries to query these data, to get their value for a specific instrument at a specific date, or to get their value in a range of two dates.
To make quantitative analysis or to write trading algorithms having these data and easily interrogate them can be convenient. The strong point of jCOT is that it's possible to make searches invoking it through shell (CLI interface) and it is therefore callable and usable by any external software or language. It can be used for example to create an indicator (graph) on platforms like MetaTrader or cTrader.
Data are downloaded from the official website here. They are saved on the disk in a folder in the format of Excel (.xsl). This allows you to use them later even without an internet connection.
- A PC with Windows (developed and tested on Windows 10) or Linux or MacOS.
- Python 3 (developed and tested on 3.10 version).
- Dependencies packages (read below to install them).
- Internet connection to download/update data (optional for a search query).
- Microsoft Excel to view (without doing search query with the script) or edit data directly (optional).
Download and install Python 3 from here. Make sure to select 'Add Python to PATH' during installation on Windows. If you use MacOS or Linux, Python 3 may already be preinstalled. In this case, check it, open a shell and type:
python --version then enter. If you have at least Python 3 go ahead. Otherwise try typing:
python3 --versionthen enter. If it works proceed, otherwise install it. From now I will write only python, otherwise it could become unnecessarily long writing python or python3, but if you noticed that you have to call the interpreter with python3, of course always do it in this way.
Click on the top-right green button 'Code' and then 'Download Zip'on this page.
Use 7Zip or WinRAR or any other archive manager to extract the downloaded .zip file in a folder wherever you want.
Open the extracted folder. Open a shell here (on Windows SHIFT + right click on a blank space and then 'Open PowerShell here' or 'Open Terminal here'. MacOS or Linux users should be experienced enough to do this 😂). Then type:
python -m pip install -r requirements.txtand then press enter. If there are no errors, the installation is complete.
There are two ways to use jCOT. The first is to simply open the program (in the extracted folder locate jCOT.py), or invoke it by shell without passing any parameters to it, so you will use the command line GUI. The second way is to invoke it by passing arguments through the shell. The latter is especially useful if you plan to write programs that will interface with jCOT.
If you are using jCOT for the first time, you will need to download the COT data. You can't research what you don't have 😂. Open the program and when asked to download the data answer with 'Y' and then enter. Otherwise, if you want to proceed via shell, open it in the extracted folder and then try:
python jCOT.py -unote '-u' which means update (or get) the data. It's recommended (but not mandatory) to have a look also at the list of available symbols. To do that via CLI GUI reply 'Y' when will be asked. To do it in a nerd way, in the shell execute:
python jCOT.py -lDuring each execution (subsequent to the first one) of jCOT you can choose between:
- View the symbols list.
- Update the data.
- Doing a search.
All these actions can be carried out through the CLI GUI, simply by opening the jCOT.py or via shell, passing these arguments:
- -u: Get the COT Data or update it if already downloaded, in this last case, it's optional.
- -l: Get the list of found symbols with the corrisponding code avaible for a query search.
- -h: To see an help page similar to this one.
- -s : The symbol code used for a query search. Required. To get this you can first call the program with -l.
- -d : The date used for a query search. Required. Format dd/mm/yyyy. You can also perform a range data query between two dates. In this case the format will be dateStart:dateStop Where the dateStart and dateStop are in the format dd/mm/yyyy. Range data queries are slow, use it only if needed.
Examples of usage with shell:
- python jCOT.py -u
- python jCOT.py -l
- python jCOT.py -u -s 099741 -d 03/04/2005
- python jCOT.py -s 099741 -d 05/06/2006
- python jCOT.py -s 099741 -d 01/01/2021:31/12/2021
Example of expected output with shell (this is the one produced by the third command in the list above, the data have been updated and a search has been requested for the symbol with code 099741 (corresponding to EURO FX) for the date 03/04/2005):
Example of expected output without shell, using CLI GUI:
jCOT takes a date and a symbol as input. It searches the reports for the symbol and returns data for the most recent date to the one entered. If, for example, for the symbol EUROFX with code 099741, on 09/02/2016 data was released and also on 16/02/2016. Entering 12/02/2016 as the search date will return the most recent data available, i.e. that of 09/02/2016. If instead you enter 20/02/2016 as the date, you will get those of 16/02/2016. in case of search by range, all data released between the two dates entered will be obtained.
The data shown are those generally considered most relevant. However, there are many others that are not shown you might be interested in. They can be viewed by opening an Excel file contained in the 'COTData' folder. Each column will correspond to a data. For those familiar with Python programming, it is fairly easy to modify the data shown in the output, by adding more data, or removing the existent ones. Get a Python code editor and open the script. Search (usually you can do this with CTRL + F or CMD + F, but it depends on the editor) for 'Modify the output'. You should find a comment with a '#' and underneath the declaration of a result (dictionary) variable. As you can guess this is the heart of what will be shown. To delete a displayed data, simply delete the related line. To add more data you can add a ',' to the end of a line, wrap and add a new line with the format 'key': value. 'key' must be a string with superscripts. 'value' (without superscripts) will be the value of the data shown. Enter as value 'resultrow["COLUMN_NAME"]' (without superscripts) and instead of COLUMN_NAME put the name of the Excel column you want to output (mantain the quotes).
WARNING: Do not modify or delete the "MarketName" and "COTDate" row since they are used by the program in the range search system.
Example:
