Trading Signatures: investor attention allocation in stock markets

Anonymised data sufficient to reproduce results in article "Trading signatures: investor attention allocation in stock markets" (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3687759).<br>open_code.py - produces the figures exactly as in the article.<br>signatures_open.csv - contains the trading signaturesportfolios_open.csv - contains anonymised information about the securities in the signatures<br>Each row in both files contain information about the sector of the investor, anonymised investor id, calendar year when the signature was estimated. In signatures_open.csv, the remaining rows indicate, what fraction of all transactions executed in that calendar year were executed with nth security. In portfolios_open.csv, the remaining rows indicate which security was the nth most actively traded security in that year.<br>Due to sensitivity of the data, investor id's in signatures_open.csv and portfolios_open.csv are anonymised independently. It is not possible to match investor signatures and their traded security compositions, however, one can keep track of individual investors signature changes.<br>effort_change.csv - contains information about changes in investor signatures, i.e., change in the length and entropy of the signature, as well as average and total trading intensities.<br>When using this data set, please cite this data source and the original article.