Skip to content
All library documents

Finding Trade-Level Data to Study High-Frequency Trading

Article Quant Q&A · Author: Dan Green-Leipciger

Summary

The document considers how to obtain transaction records for research into high-frequency trading’s effects on individual firms. It points to commercial data providers as a source of tick-level transactions for selected securities and exchanges, with coverage that may extend back several weeks. Researchers may also find longer historical datasets held by organizations that collect market data for analysis and strategy testing.

The discussion emphasizes practical limits: comprehensive trade and market-depth records can be very large, computationally demanding, and costly to maintain, so extensive proprietary datasets may not be readily shared. It suggests contacting a trading firm to ask whether it can provide anonymized records. The exchange offers no procedure for identifying which firms were algorithmically traded from public data, and it does not validate a method for making that classification. The proposed regression is mentioned as the researcher’s plan, not as an analysis performed in the document.

Key ideas

  • Commercial data vendors can provide tick-level trade records for selected securities and exchanges.
  • Historical market datasets may exist at trading organizations, but their availability to outside researchers is uncertain.
  • Trade and market-depth archives can require substantial storage and computation.
  • A researcher could ask a trading firm whether it will share anonymized data.
  • The discussion does not establish a reliable way to label firms as high-frequency traded.

Tags

Full text
# How to obtain a log of all trades done on the Nasdaq or other major US exchange?


# How to obtain a log of all trades done on the Nasdaq or other major US exchange?












I'm looking to do a research paper on the impact of high frequency algo-trading on individual firms.

In order to do that I need to be able to determine firms that have been high frequency traded.

My plan is to find raw trade logs, and parse them looking for patterns which indicate that they are being algo-traded.

I have so far been unable to find these logs for any major US exchanges

Does anyone know where I can find these logs?

* Edit *

I will ultimately be doing an OLS regression to determine what affect being algo-traded had on these firms

If anyone knows another way to determine if a security has been algo-traded that would also be greatly appreciated

## Answer by Eric (score 3)

https://quant.stackexchange.com/a/16272

If you're looking for all transactions against any or a given set of securities on whatever exchange, you can get that from a data provider like IQFeed or eSignal. Most of them will have tick level data going back for at least several weeks.

Some people have been collecting tick and market data for quite sometime against a variety of securities, and as such have pretty vast databases that are used for data mining, analytics, and the testing of new strategies. Some of these databases go back several years.

Those data sets are quite large, computationally intensive, and a big part of the profitability for that organization. I really can't imagine anyone parting with them, and we're talking about ~10GB a day depending on how much data you want (market depth, which exchanges, which securities, etc). So easily into the multiple TB of data range.

## Answer by emcor (score 0)

https://quant.stackexchange.com/a/16269

I dont think any of the algo traders publish their logs, you should contact an algo trading company and inform them that you would like to write a paper and ask them if they can provide you some anonymized data.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.