good

The data set includes trades from
orients between the middle of March and early June 2023
We want to better understand how clients trade and identify those that are not trading their full interest.
Analysis project: For each client, compute the autocorrelation in the direction of their trades per currency pair. Clients with highly auto-correlated trading directions are probably clipping. Use Python to complete the analysis.
The data to analyze can be downloaded from the link below. It is a very large data set with 3,000,000 rows. You may want to test your python code with only a subset of the data, and then run it on the full data set after you are sure the logic is correct.
https://barclaysnam-
my.sharepoint.com/:u:/g/personal/kevin chipalowskybarclayscom/EScTfWO3LR1Oukjaofki3g oBIFodZWgVg1|HpC3neX6VHw?e=6GCVd6
The data is a CS table with one row per historical trade. The table includes these columns per trade:
• groupNance: The Barclays identifier for the client who traded, as a text description. groupld: The Barclays identifier for the client, as a unique number. sym: The currency pair traded.
tradeTimestame.
IbsFlow; The signed amount traded from the perspective of Barclays. For example, if the number is -200 then Barclays sold 200 of the base currency and the client bought 200.
• tradeld: The Barclays identifier for the trade, as a unique number.