A popular class of strategies in the foreign exchange (FX) markets involves time-series momentum. It is based on buying the currency pair that shows recent positive returns and selling those that demonstrate recent negative returns.
Cambridge Square Capital LP (Boston) provided a pickled dataframe of USDJPY (Japanese yen) and USDMXN (Mexican peso) exchange rates.
A memo to Cambridge Square Capital recommends a simple (5, 20) moving average rule assuming active management of the pairs on a monthly basis. The supporting analysis is included in the Jupyter notebook.
All analysis was completed in Python 2.7 using Pandas. To run, create a conda environment and install the requirements:
conda install -c synthicity prettytable