Fulvio Corsi, Stefano Peluso, and Francesco Audrino, "Missing in
Asynchronicity: A Kalman-EM Approach for Multivariate Realized
Covariance Estimation", Journal of Applied Econometrics, Vol. 30,
No. 3, 2015, pp. 377-397.

The data used in this paper are part of the NYSE trades and quotes
(TAQ) database. Since the NYSE does not permit open access, the data
cannot be stored here. TAQ data are delivered via FTP but can also be
accessed by connecting to the NYSE Technologies Market Data Analytics
Lab using either a web-based interface, Microsoft Excel plug-in, or
API. For additional information, please refer to the Historical Data
Products section of NYXData at:

  http://www.nyxdata.com/Data-Products/Historical-Data

The analysis in the paper is based on data for 2006-08, distributed in
roughly 50,000 .mat files and 100 folders that occupy about 1.5
Gigabytes. After ASCII conversion and zipping, the space occupied is
roughly the same.

Each folder corresponds to a financial stock analyzed, namely:

'AAPL', 'XOM', 'MSFT', 'IBM', 'GE', 'CVX', 'WMT', 'T', 'PG', 'JNJ',
'WFC', 'JPM', 'PFE', 'KO', 'GOOG', 'PM', 'ORCL', 'INTC', 'MRK',
'QCOM', 'CSCO', 'VZ', 'C', 'PEP', 'BAC', 'MCD', 'COP', 'ABT', 'SLB',
'AMZN', 'BRK.B', 'DIS', 'OXY', 'HD', 'UTX', 'CAT', 'AXP', 'KFT', 'MO',
'CMCSA', 'V', 'MMM', 'EMC', 'UNH', 'GS', 'USB', 'CVS', 'AIG', 'UPS',
'BMY', 'BA', 'AMGN', 'UNP', 'MA', 'DD', 'EBAY', 'HON', 'HPQ', 'LLY',
'CL', 'F', 'SPG', 'MON', 'SBUX', 'ACN', 'DOW', 'MDT', 'NKE', 'MET',
'APC', 'COST', 'TGT', 'LOW', 'MS', 'DHR', 'SO', 'APA', 'TXN', 'EMR',
'GILD', 'TWX', 'FCX', 'PCLN', 'PX', 'PNC', 'CELG', 'NOV', 'DTV',
'EXC', 'BAX', 'NWSA', 'YUM', 'DE', 'COF', 'HAL', 'BIIB', 'EOG', 'PRU',
'TJX', 'DELL'

In each folder, a .mat file corresponds to a trading day. For
instance, AAPL\20060103.mat contains all the tick-by-tick prices and
times of the trades for the stock with ticker 'AAPL', in the trading
day 3 Jan 2006.