aggregating statistics

Discussion in 'Technical Analysis' started by saminny, Jun 20, 2008.

  1. saminny



    I have accumulated the following data for a set of securities I have been working on for from different sources. The data that I have is computed daily for each security and each exchange. So I have data separate for IBM,NYSE and IBM,ARCA and IBM,BATS..

    1. No of changes in level 1 bid/ask.
    2 Average quantity at the top of the book.
    3 Avg quantity weighted by time at the top of the book.
    4 Avg Spread
    5 No of trades
    6 Avg Trade Qty
    7 Avg Trade Price
    8 Volume traded
    9 Avg Level 2 Book Qty (+/- 20 bps from midpoint)
    10 Percentage of time the exchange has best bid/offer for that security
    11 10-min volatility (std dev of returns every 10 mins)

    Now I am looking to aggregate this numbers across certain periods, so that I can present a 20-day (or monthly) moving average of these statistics.

    I need some suggestions in the best way to aggregate these numbers that will make the most sense and add more useful information.

    for average trade price, I am aggregating by taking a volume weighted average price across days.

    Any suggestions, comments and observations are most welcome.