What's in the Session?

In this workshop, OneTick Senior Cloud Architect, Mick Hittesdorf, will examine how cloud-delivered, managed market data platforms are reshaping the build-versus-buy decision for quantitative research infrastructure. Our audience will better understand the practical trade-offs, architectural considerations, and operational implications of moving from self-built data pipelines to on-demand, pre-normalized data environments.

Agenda:

• The Data Infrastructure Bottleneck

• The True Cost of Building In-House

• The "Buy" Solution: Enter OneTick Cloud

• Matching the Solution to Your Fund

• Under the Hood: Institutional-Grade Quality & Access

• Solving Real Problems Quickly: Live Demo

• Next Steps & Q&A

Watch the Recording:

Webinar Transcript:

Thank you for joining us today. I'm Justin Fry and I'll be the host for this session. A few quick housekeeping notes before we dive in. We'll hold a few polls throughout the webinar and there will be a dedicated Q&A at the end.

Please feel free to drop questions into the chat at any time. And thank you to those who submitted questions ahead of time. We're delighted to have Mick Hittestorff with us today. Mick is the Product Architect for OneTick Cloud where he leads technical architecture and strategy.

He's a hands-on technical leader with a strong practitioner background, previously serving as the head of US Data Engineering at Akuna Capital and holding senior data and engineering roles at Transunion, Chicago Trading Company and ABN AMRO. Thanks so much for joining us. Today we're tackling a build versus buy dilemma.

Thank you so much for that introduction, Justin. And thank you to everyone who has joined our webinar today. I want to start by acknowledging that your time is valuable and we recognize that. So it's my goal to leave you with some actionable information and next steps to tackle what can be a very problematic and difficult dilemma to address, to build versus buy, which is why we're all here today.

By way of introduction, Justin did mention that I have extensive experience in the data and trading industry, and I want to share with you some recent experience just a few years ago where I led a data engineering team that actually built an in house data platform, market data platform, ingesting both historical and real time data. So this knowledge is not theoretical or conceptual… I've been in the trenches, and want to share some hard-won advice. I have been a builder, and now I'm on the other side of the fence where I'm gonna tell you about the advantages of buying and what OneTick Cloud can do for you and your company to accelerate time to value.

So without any more delay, let's get into it. So this will be the outline. Again, your time is valuable. So I will spend as little time as possible just setting the table around the OneTake cloud value proposition and then just some of the challenges and costs associated with an in-house build versus what we think is the superior alternative OneTake cloud, the market data and the analytics that OneTick Cloud provides. So I'll go through, you know, what data infrastructure is and, you know, what you typically need to do in order to have a robust data platform that supports both historical and real time data as well as analytics.

Really, what is the cost of building in-house? We'll cover that briefly.

The buy solution, which is OneTick Cloud, and and I'll review some of its capabilities and why we think, you know, that should be your first choice.

And I'll also help you based on your persona or your institutional profile, what capabilities of OneTick Cloud might map best to your own individual requirements. And then I'll actually talk a little bit about the data. The most important component of OneTick Cloud of course, is the data. It's institutional grade, high quality data, and then how to access that data via our OneTickPy Python API. And, you know, by that time, you should probably be, you know, tired of slides.

And, you know, for those in the audience who might be more technical and might not be convinced by slides but would rather see working code, I have a Jupyter Notebook with a number of examples, progressively more sophisticated examples that demonstrate the basics of using OneTick-py through more intermediate and then some advanced examples. And then I'll leave you with a suggestion to do this yourself. We offer a free trial account with OneTick Cloud. I can go immediately. I'll show you this page again at the end. You can go immediately to this page. You can create a trial account, and we have some trial databases that you can access with real data, t plus one historical data.

You can download and install via PIP, our OneTick-py API. It is a free open source MIT licensed API. That's the API library that I'll be using as we go through the examples today. And I would encourage you to, you know, do the work yourself. There's nothing like hands-on learning to give you a real feel for the power and capabilities of a OneTick Cloud.

So that's the organization and what we'll cover today.

So what is the data infrastructure bottleneck? Right? So if you are a fund or a systematic trader, maybe a pod in a larger firm, or maybe a small even a small family office, you know the value of data and how difficult it is to obtain high quality data. That's a prerequisite for doing anything.

So very often, if you're a PM, a quant, a trader, maybe even in compliance or doing some sort of reporting, the first thing you need is high quality data. And that becomes your first build versus buy decision. Where am I gonna get that data? Am I gonna buy it from a vendor?

Am I gonna buy? Do I need real time or just historical? Do I need both?

Once I obtain that data, how do I know that it's good data? Where am I gonna store that data? How am I gonna provide access to the data through what APIs, through what platforms, what tools? There's a tremendous number of decisions that go immediately into the decision to enable your users with a data infrastructure that can meet their problems. And historically, many surveys have been done and a lot of analysis has been done.

What's been concluded and is really without argument is that Quants, for all their sort of intellectual capabilities and skills and the cost that goes into hiring these very expensive people, we'd like to think that they're building alpha generating models, right? That are making the firms that they work for money.

Turns out that most of the time, because again, clean high quality data is a prerequisite for a high quality model or analysis or a trading signal, that the first thing that needs to be done is to validate, you know, curate, clean, align, and reconcile the data itself. So eighty percent roughly is the number that historically Quants and other people who work with data spend actually cleaning data. So that's eighty percent of their time spent wrangling data before they even get a chance to do the high value work that they're paid to do.

And then lastly, let's say you have the data, how are you going to quickly derive value from that data? What tools do you have? What's your stack? There are a lot of questions there. Do you go open source? Do you go all in on a vendor's, you know, siloed proprietary stack?

Do you install that on prem or do run-in the cloud? And there's all sorts of variations of those alternatives that need to be considered. So at the end of the day, because these are all complex decisions with a lot of dependencies, it takes forever to get reliable, clean data for one market and much less multiple markets, especially let's say you're a multi strat or a global macro fund where you need data from multiple exchanges you know, different countries.

So what does it take to build a market data platform in house?

My experience at Akuna Capital, where I live the data engineering team, is it takes six months typically just to get to to an MVP, to a data platform that might support one exchange, you know, in a couple protocols or a couple datasets, maybe an order entry exchange and a market data exchange.

And then it takes many months more as you add and onboard more exchanges, more pipelines, etcetera.

It can easily take, you know, two plus years to get to a robust data platform with meaningful data. And of course, there's no history then in the system. Once you start, you start with nothing. Right? Maybe you have some PCAPS, but then those need to be loaded as well. But very often, then you have no history at all. And then and then if you wanna integrate a real time feed, well, that becomes an order of magnitude even more difficult dealing with reliability, latency, failover recovery, etcetera.

It takes quite a bit of sophisticated engineering also to integrate real time and historical feeds and and and reconcile schemas and timestamps and things like that.

So, you know, breaking it all down, you're gonna spend at least two months, eight weeks, you know, deploying infrastructure, another couple months sourcing, loading, and validating tick data, cross referencing, you know, you have one symbology or multiple. Maybe you wanna use your users wanna use Bloomberg, ISENS, QSIPs, maybe an exchange identifier they might be familiar with, etcetera. So if you even do it or have time to do it, getting cross referencing symbols might take you another couple of months, another eight weeks. What about if you're, you know, trading equities?

Right? And you absolutely need to apply corporate action adjustments, and I'll show an example of that when I get into the code, how important that is and what a difference it makes in the actual data, you know, if corporate action adjustments like stock splits are not applied. So that's another few weeks, maybe a month. Right?

And then what about in the US, like, where there's sixteen SEC exchanges with liquidity across multiple exchanges, plus a SIP. Right? So are you going to integrate and consolidate and cross reference all that data the different time stamps, the different exchanges, the different condition codes, etcetera.

Just integrating all those exchange feeds into a single consolidated consistent feed that could be analyzed in a way that it would be suitable for building a trading signal or feeding a machine learning model. It's gonna take you another couple months, most likely. And then, who do you need to build out this system? 

A lot of hard problems to be solved that typically require very highly skilled, expensive engineers. On my team, I had three to four sometimes more highly skilled, very expensive developers dedicated full time to these tasks. And, again, it took us six months just to get to, you know, having a working platform that supported exchanges feeds and then, you know, another two years approximately to onboard another eight or ten exchanges. 

Then, of course, that doesn't account for all the maintenance exchanges. Specs change. There are new regulations. The SIP changes, for example, recently went to, you know, decimalized or fractionalized orders. So now you can get orders with quantities of point seven five or something like that. So what if your feeder can only handle integer trade sizes? Right? Well, then all your pipelines break.

We handle that in OneTick Cloud.

And then you need someone who can do support administration, maybe DevOps, data operations team to monitor loads, data quality engineer perhaps to write tests and follow-up on failed tests, etcetera. So very quickly it becomes a very expensive proposition to build a market data platform in house of any scale and scope.

So what's the alternative? Right? I've been on the build side of the fence. I know what's involved.

The superior solution is in terms of both capabilities, data quality, and obviously, most importantly, time to value is, you know, a cloud-based buy solution. So that is OneTick Cloud. It's a you know, our OneTick Cloud platform, you know, is fully managed. The automatic updates, the infrastructure, we load the data every day.

We have a partnership with a real time feed vendor, Options IT Active Financial, where you can get real time data through OneTick Cloud.

And that helps you to shift this very capital and heavy CapEx heavy proposition and project to a predictable vendor managed OpEx expense model. And then you get instant access to datasets. We have over two hundred global equity futures and options databases plus FX.

Recently, we added some fixed income, some rates information, as well as crypto, like, for example, from Binance and a few other crypto exchanges. And the cloud environment by definition inherently, right, is expandable. We can assign to our users, to our firms, the capacity that is required in terms of cores and memory, and it's very easy to, you know, increase, you know, the allocated compute capacity if necessary.

So just to summarize, you know, put it all in one table. You know, you're looking at, this is an estimate roughly, but thirty four to sixty weeks. So really talking about more than six months to potentially over a year to accomplish all these tasks that I went through.

With OneTick Cloud, of course, there's no pipeline and database development nor tick data loading and cleaning. We do that for you. So it's really not even applicable. But as far as corporate action adjustments and symbol mapping and continuous futures and then bar generation and consolidating liquidity, these are all hard problems.

Very often with an in house development project, that project team never gets past one and two, building pipelines, building databases, and loading and cleaning data. Things like corporate action adjustments and symbology mappings and continuous contracts and building real time bars or even two plus one bars, etcetera, or normalizing across different exchanges. These are all extra credits. Right? Things that the team, you know, hopes to get to, but often end up being a lot of Jira tickets in the backlog.

So, you know, we've solved all these hard problems. You can now take advantage of the work that we've done, that, you know, we have done for, you know, dozens of clients already.

Okay. So matching the solution to your needs. Right? So who are you? Right? Maybe you are a brand new fund.

Right? You have next to no technical team in house.

You need data, as I mentioned, to do anything.

You have maybe a strategy that that you feel has a a good sharp ratio, for example, and you're you're eager to deploy capital, but you need to validate, let's say, historically over some period of time, you know, and you need maybe you need both, you know, US and international equities data or maybe you need some options data or you maybe need some foreign exchange data, whatever the case may be. All that data is available, you know, again, over two hundred exchanges, over twenty years of data in some cases, depending on the asset class.

So rather than spending six to twelve months, you know, waiting to even launch your fund and deploy capital, you can immediately validate your strategy using our data and hopefully start generating returns for your investors. For more established groups, let's say a multi strat or or, you know, multi manager fund with multiple pods, it's kind of similar. If you're starting a new pod, you're kind of in the same place as a new fund.

But let's say you are in a firm that does have infrastructure, or there may be legacy databases with, you know, data that is of dubious quality.

You may have multiple market data vendor contracts that you're managing, which can be expensive and difficult to, you know, to establish compliance, you know, with the terms of those contracts.

And you may have multiple compute silos, different database vendors, different data in different formats, etcetera, all of which makes it very difficult to move quickly and to get access to the data you need. I mean, you may be, very familiar with, you know, the need to go to a central data engineering team or or or market data management, you know, group and, you know, maybe you have to submit a ticket or something to get access to the data you need, and that might take, who knows, weeks or months. So, again, the alternative is, you know, subscribe to the data in OneTick cloud that you need. Even for very focused use cases, you can subscribe to just the symbols, just the history, just the asset classes, the exchanges you require in your pod or for your strategy and immediately get to work.

And do so using, again, a high quality, easy to use, intuitive, unified, you know, Python API, which I'll show in a bit.

So and then and then, you know, lastly, just to be, you know, very specific, I mentioned, you know, model development and a m AIML feature engineering. You can also do, of course, microstructure analysis using our order book order book analytics capabilities of OneTickPie. Other common use cases include execution analytics, you know, BestX. Let's say if you're a broker and need to establish compliance, you know, with applicable, you know, execution regulations, that's a common use case.

And because we have all the trades and all the quotes, you know, within OneTick Cloud that are inputs into a Best Ex analysis. And or maybe you are managing a fund, you just need to make sure that you're getting the, you know, the best fills, you know, so you need to calculate VWAP, TWAP, etcetera, and that's easy to do as well within OneTick Cloud. So all these use cases and others can be you know, are supported by OneTick Cloud.

So one of the things that I like to emphasize is that the data in OneTick Cloud is not, you know, retail data. It's, you know, our core customer segment is, you know, our institutional customers. We have some OneTick, you know, has been serving, you know, Wall Street and, you know, professional fund managers, brokers, market makers, exchanges, you know, our our customer base plus now the, you know, the KX customer base that, you know, KX having, you know, recently, and OneTick recently merged. You know, we have the, who's who on Wall Street as our customers. So we take data quality very seriously, and we understand the demands of, you know, these institutions.

We've demonstrated for years that we understand what institutions require and what institutional grade data is, and we're confident that we can serve, you know, the needs of institutions. So, you know, it's not just an end of data. You can access tick by tick data, full depth, data with L2 and L3 depending on the exchange, depending on the asset class.

And the the engine underlying everything in OneTick Cloud is our industry leading high performance time series engine, OneTick time series engine that allows you, you know, you to do very easily, joins between ticks and quotes, and other sorts of analyses, which I'll demonstrate in just a few minutes.

And then in addition to the the Python API that I've already mentioned, we also support standard SQL, that's what you prefer, REST, and then also, optionally, you know, you can work with us to receive on a regular schedule bulk CSV or Parquet downloads, which is something we do, again, do quite often on the t plus one basis so that our customers can, you know, backfill or augment, you know, their data with onto cloud data.

Okay. So now let's kind of set the table.

I like to see code, and, hopefully, some of you would also like to see some code. So I'm gonna go to a notebook which I put together that basically introduces and describes how to use the OneTickPy library, which is our Python library. OneTickPy is an open source MIT licensed library. There is a you know, at the end of this document, I provide a link to the GitHub repo. So you can actually go to the GitHub repo, clone it, all the source codes there. There's no surprises for not hiding anything.

One of the primary virtues of the OneTickPy library is that the API is consistent with pandas. So pandas, for those, you know, amongst you who work with data and as part of you know, for you know, professionally, You know that Pandas is the has been for years the the leading data analytics library in Python. It's a data frame library.

There's a huge user base.

Everyone from, you know, undergraduates to graduates to, you know, professionals who've been in the mark know, you know, working for, you know, twenty plus years have have probably used pandas at one one point in their career or perhaps even experts in pandas and prefer to work with pandas data frames. So what we've done is basically taken the pandas API and grafted it onto OneTick. So this is, you know, the standard Python API for the OneTick database, but also since the OneTick database is what is used to power OneTick Cloud, that is then the standard, you know, Python interface for OneTick Cloud and accessing our data and accessing our analytics. So in particular, I wanna emphasize that if you're familiar with the kind of definitive book, Python for Data Analysis, which you, you know, may or, you know, have seen, you know, in the I'll put it here.

You know, maybe you have read yourself and consulted those examples and learned how to do time series analysis, then you'll recognize the simplicity and power of doing, you know, time series analysis with pandas. All those same operations plus many more are available through OneTickPy.

So, again, you can do pandas. You can write panda style code, but you can go way beyond what pandas can do in that the actual operations are actually executed in the OneTick database. So we can handle much larger time series than you can do with pandas. We can do it faster. And then we've built a lot of custom analytics on top of OneTickPy, which I'll show you in a minute, that allow you to do much more than you can do with pandas alone. So, you know, to get started, you know, I would recommend, you know, just following these instructions, which are available on our you know, on the OneTickPy documentation, which is online. You can just Google OneTickPy, and you'll be able to find that documentation.

And I would also suggest that you sign up for a free trial account for OneTick Cloud. So I, of course, already have an account, but if you come in here and then go in once you've signed up, you'll be able to go into OneTick Cloud, which is kind of what this is the landing page. And then from here, I can begin seeing everything that's available. I would recommend starting with data assets, and you can then see you'll have availability or access to our sample databases initially. Those sample data databases, you know, are permissioned to trial users. Right? So if you wanna look at, you know, US Equities or CME, EUREX, etcetera, it's all here.

This is probably the best place to start would be a US comp sample. So this is all of the US SEC exchanges, all of that flow consolidated. It's tick by tick, and you can actually see the data that's available. So these are both the tick by tick database and the bars database. These are one minute bars. So open up, open high, low, close.

And then these are the tables, which we are in OneTick Parlance called tick types, but you can just think of them as tables. So for example, if you wanna see quotes, right, we have a quote table with all these fields, NBBO, right, as well, trades with all these fields. Right? So this is kind of the standard structure that you would see. And, again, if you once you sign up for a trial account, you'll have access to all of these sample databases.

So let's come back here.

Also in OneTake cloud, which is important not to under underemphasize, is reference data. So things like exchange calendars or exchange sessions, pre market hours, regular trading hours, post market hours per exchange, mappings across symbology. So OneTick Cloud, as I mentioned, supports multiple symbologies such as Bloomberg tickers, ISINs, QSIPFIGI, which is the Open Bloomberg symbology standard, as well as corporate actions and adjustments.

And, again, you can access all of the OneTick Cloud reference data via the OneTick API, which supports both Python and SQL.

So I'm gonna cover a few examples.

I'll cover some basic examples, some intermediate examples, and then a couple advanced examples. The last example, won't I won't run. I'll tell you the code.

But all the first five, I'll show you the results of these.

So let's look at, you know, the simple case of querying trades and quotes. Right? About as basic as you can get, and you can basically get access to trades within, like, six lines of code here. Right?

So in this case, I wanna query Apple. I specify the date, which in this case is April thirteenth at ten o'clock this year. I wanna look at one minute of trades. Right?

This is the name of the database. The tick type, that's what I said. That's your table. So you're basically looking, you're querying the US comp database and the trade table, and you basically call this method called run.

Right? We're passing the query in this case, which is just this data source and the start and end time and the symbols you wanna see. And then in return, you'll get all these you'll get your trades. Right?

So we have the time stamp, the exchange time stamp. So the TRF time, that's basically, you know, when the consolidator the SIP, for example, time stamp that.

The source exchange, like I said, is consolidated, so we have exchange codes, price, and size. Again, I wanna emphasize that we already support fractional sizes within OneTick Cloud.

Again, this would probably break a lot of in house pipelines. I'm sure there are a lot of teams that scrambled or are still scrambling to support, you know, fractional trade sizes, and we've already supported that for you. Not nothing for you to do. And a lot of other fields.

Within the OneTick Cloud dashboard, you can find, you know, essentially a data dictionary that describes all these fields if you're curious.

Similar code for quotes.

Again, if I want… if I wanna see the national best bid and offer, it's just, you know, querying the same one minute of quotes is, you know, extremely easy. In this case, you can see, you know, two lines of code to get access to quotes.

So now you've got trades and you've got quotes. Right? So let's say you wanna join trades and quotes, which is a very common operation, but it's actually quite difficult to do with a lot of traditional databases where often the best you can do is kind of use windowing functions or something is try to line up trades and quotes.

Being a time series database, it is a primitive operation within OneTick to join trades and quotes. So let's say I wanna join, you know, again, Apple that same, you know, one minute. So if I wanna join, you know, join all of the trades, right, to the corresponding quotes, you know, you get this combined data frame which shows you both. Right? So for a particular trade, you can see exactly what the bid and the ask and the size was for that.

Again, a couple lines of code and that work is done for you.

I mentioned corporate actions. Right? So here's an example of what looks like a very odd chart. Right?

So looking at daily prices for WMT, I think this is waste management, on this day, which was, you know, February second and February third. So I think a twenty four hour time period is 2024. What was special about this day? Well, there was a three for one split or a one for three split.

And so if you didn't know that happened, you would think, well, my gosh. There's some really bad news or something. The stock price just crashed this day. If you tried feeding this raw data into a model or an analysis, right, this would totally corrupt your model.

Right? Make or and, you know, make any results related to this trade or maybe even this whole day if you're looking at volatility or something to be completely unreliable. Right?

So, you know, you need to adjust in order for this data to be usable, you need to adjust for this split. And, in OneTick Cloud, we make that very easy.

Again, you know, less than a dozen lines of code, We have the ability to apply corporate actions to both, you know, prices and sizes.

And then here's your adjusted, you know, close, you know, price here.

You know, the original was, you know, this case was around one seventy. You know, it was a three for one split, so now it's, you know, fifty six point four two.

Actually, one for three. And then if you plot the actual adjusted prices, this looks a lot more sane. Right? This is kind of what you would expect, that crazy what looked like a drop in price has been eliminated, and you have, again, a reasonable high quality time series that can be used for analytics or for building models or signal generation.

Again, that was done with just a few lines of code. And that involves not only the code, but actually having access to things like adjustment factors and when the corporate action was actually announced and when it actually, you know, took place on, you know, on what trading day. All that reference information is available in OneTick Cloud.

And so let's go on to some slightly more advanced analytics, which would be querying order books, for example. We have a lot of very sophisticated and powerful order book analysis functionality.

And so in this case, I'm showing an example of querying a full depth book. So this would be a market by order level three book at a specific specific time to a maximum number of levels, which you know, so you can specify how many levels you wanna see. And in this case, I've specified that I want five levels, and this is gold, ComX gold futures.

So I base and so one nice thing about the databases as well is we it's easy to query all the symbols. So if you wanna see all of the symbols, in this case, I'm just using a regex to find the one ounce gold future any month in twenty twenty six, printing those out here. So I can see, you know, June, you know, through December, those gold futures for the one ounce gold ComX contract.

And then well, what's the front month gold future? Well, it's the first one. Right? So in this case, it's the June twenty sixth future.

PRL full, that's the actual tick type or database name. This is, you know, price level full depth.

I wanna see the five levels of the book, and what you see then is this here. Right? So and and then one is basically see the bid and the ask, the size. Right? So, you know, we need to differentiate between bid and ask to the zero and and and one.

And then if, again, if you just wanna see what the fields are in your full depth book that's documented in the OneTick Cloud dashboard, but I put a screenshot in here for convenience.

And then the last example, which I'm calling an advanced example, not because it's hard to do, but maybe, you know, calculating time weighted book and balance is is a concept that's not familiar to you or it it might be difficult for you to calculate this to get the data lined up in terms of time correctly, aligning time stamps, etcetera.

What books, you know, are you reconciling, etcetera. So there are some considerations that go into, you know, calculating a time weighted book and balance, but again, we make that easy. Right? So we have this we have this very powerful but very easy deceptively easy function called OB snapshot wide, which will basically take a snapshot of the book in this case rather than what I showed in in the book kind of stacked on top of each other, this shows you the bid and the ask next to each other horizontally. And in this case, I'm showing just three levels.

And so you can see the for this particular day, which is January thirtieth, for a hundred milliseconds, you know, point of point in time, what the book looks like for gold futures, including, again, the bid, the ask, the bid size, the ask price, and the update times, like when that actual quote hit the book. And then this code potentially does some aggregation and computes the imbalance across a hundred and twenty five milliseconds. Right? So as long as quotes are hitting the book and updating the book, you can see buy time, the ask volume and the bid volume, and then the corresponding and balance. Whether it's, the book is leaning towards the bid or leaning towards the ask. Right? Which for micro market microstructure analysis, can be very insightful.

And then the last example, which I'm not gonna show, but I just wanna kind of leave you with a teaser. We do support real time processing using our active partnership with Options IT. This is a premium feature, but there are those firms that want real time data and for its historical data isn't or even t plus one data isn't sufficient. So we do support processing real time data using a callback pattern.

This example just shows kind of a classic simple vanilla example of a golden cross where you're looking at basically two moving averages of different durations and looking for essentially when a long term moving average crosses is the short term moving average or vice versa. And a lot of technical traders will use that indicator as a signal to either, you know, buy or or sell when that golden cross occurs. So very simple to write this code. We use a callback mechanism. Essentially, what you do is you register this query with OneTick Cloud and as tick by tick as the real time market data enters OneTick Cloud, the cloud will call your code on your computer, and this code will be evaluated.

And so I wanna leave you with that. I know I've covered a lot of ground, but I hope that I've been persuasive that it's very easy to do nontrivial, you know, operations and analytics with OneTickCloud, and we provide a tremendous breadth of data, whether it be, you know, US or international exchanges, multiple asset classes, futures, equities, FX, crypto.

Again, all accessible through a unified Python API, which is open source.

So I'm gonna wrap it up there because I know your time is valuable, but I wanna leave you with this.

You don't need to make a huge investment in your own market data infrastructure nor do you need to make a huge investment in time spent learning how to use OneTick Cloud or even using it in production within your own firm. We can work with you to provision, you know, as much data as you need. You can start with a small dataset, whether that be just one market, US equities or a particular country like Taiwan or India equities, particular asset class, maybe you just need FX data or or, you know, CME futures and options data, we can do that. If you have a specific use case, we're also you know, I've showed you a lot of examples, but we have experts in capital markets, professional services, people on and people on the product team that, you know, are happy to to talk to you about your requirements and, you know, tailor, you know, the OneTick Cloud and what it can do to meet those requirements.

And I always encourage you to before you pull the trigger on an expensive in house development project to at least evaluate OneTick Cloud. You can get a free account. You can download this free open source Python API. So friction is low.

The cost of entry is extremely low.

We'd love to talk to you.

So start a trial account, start looking at the data we have, install OneTickPy, PIP install into your own environment, into a notebook or, you know, your Python IDE of choice, and try running through some of those examples yourself.

And I think you'll be happy with the results. So, again, if you have any questions, please reach out to us. We really wanna help you do things faster, easier, and we wanna help you accelerate your time to value.

So I'll leave it there. And, again, thank you very much for your time. I hope you hope your time was well spent.

So the question is, does it make sense to have multiple vendors for tick data?

Yes. It can. We do see that quite often. It's counterintuitive that some people think, well, I have my data. I don't need anymore.

Many of our customers do in fact source data from multiple vendors.

Data can vary in its quality, in its coverage, in the history that is provided, in the latency in which it's delivered, and most importantly, in its quality and its structure, kind of the schema that it's you know, in which it's delivered or even the format in which it's delivered.

So data is not one size fits all. So just because you, you know, have one or maybe even more market data vendors internally, you know, that you know, those vendors may not be meeting your needs, And it just may be difficult and time consuming, as I mentioned earlier in the call, to actually get access to that data. Especially in large organizations, There may be certain structures in place, processes, etcetera, that require you even internally to jump through some hoops to get access to the data you need. And so we're just trying to shortcut that process and make it easy to get data. And then another use case is beyond just ease and and and time to market is is backfilling data that you maybe you do are collecting data, but you want an independent objective third party source to use as a backfill source in the event that your primary feed fails, or you want to compare and reconcile your primary feed with a secondary feed, let's say to fill in gaps, for example.

And then lastly, it's a very common use case. Let's say a firm, a trading firm is evaluating entering a new market and wants to get let's say they wanna evaluate, you know, market making in Brazil or something like that, Bovespa or in NSE, right, or in some other exchange or or even a new asset class. And they don't even have that data on-site. They don't, they're not collecting it.

They could be a very sophisticated firm that does a lot of their own data collection, but are not yet, you know, willing to invest in, let's say, a colo or other sorts of infrastructure at a new exchange. So it's very cost effective to purchase the data from us, right, or you know, and do that analysis. Look at the book. Look at the quotes.

Look at the ticks. Look at the volumes, etcetera, liquidity. So those are all really good use cases where even if you have in house data, then it makes sense to still go to another vendor for that data like OneTick.

Thanks, Mick. The question, how can a charting website slash platform combine cloud infrastructure and delivery in combination with an existing proprietary database?

We do that and support that use case through the OneTickPy API. So you can basically run that API in in our case, you know, from something as simple as a notebook and combine that with, you know, local data that you have or embed that API into a pipeline or an application. And then in that case, you can get market data, whether it be t plus one or real time from OneTick, and then also access in that same application, that same pipeline, that same notebook, your own in house data. So that that's a quite common use case, especially for, like, TCA where you have your own trades locally, your own transactions, and you wanna join those with quotes that are available the cloud and perhaps even compare some of your own fills to, you know, market trades that are also available in our cloud. So that's kind of typically how you would accomplish that task.

Got a couple more questions. I think it requires a call to the individual who submitted them, but I'll put them to you, Mick. How do you evaluate the different vendor options in the marketplace?

You're gonna be looking at all those dimensions I mentioned earlier in the call. Right? Quality of the data, I think, is extremely important. Coverage in terms of the data itself.

In terms of the technology, you may be looking at whether it's proprietary open source. What formats do they support? Right? So OneTick recently added support for Parquet for its actual native archive format without compromising performance.

So, you know, OneTick can now support a more of a you know, an open interoperable data lake strategy that may be important to you, important to a lot of our customers. So there it kinda depends on the dimension. You know, those are some of the considerations that I would definitely take into account when doing any sort of build versus buy or build versus buy evaluation.

Thank you, Mick. We'll be in touch with the person who sent us that question. Another question here. Level three streaming data is timestamped at what interval?

That's where I would probably need more context. Depends on the asset class, depends on the actual exchange. Typically, there's not just one timestamp. There's multiple timestamps. You're gonna get typically the exchanges time stamp in the feed, and then the the capture time stamp is gonna be another time stamp, which is useful, you know, to actually kind of it kinda measure the the hop and that the time it takes to get from the exchange to the actual colocation site where the data is being collected. At least those two timestamps and in some cases also, like for a SIP, you're going to see the SIPs, the aggregator's timestamp as well. So those are some of the timestamps that we collect right now in one tick.

A typical question we get from someone interested in data is, how's your data compared to big distributors such as Bloomberg, FCG, FactSet in terms of cost, coverage, and quality?

I think the answer depends on, you know, which one of those vendors. You know, some of those vendors are notoriously expensive and have extremely strict requirements in terms of redistribution. You know, I don't wanna bad mouth any vendors, but there are vendors who make it very difficult to use data extensively. We don't have any of those constraints. You know, our constraints are, I would say, relatively minimal.

The quality of our data is something we take very seriously. We have an extensive operation staff. We have recently been announcing in the near future a commercial data quality platform. It's something we use in house, and we are commercializing that.

So you'll be able to use that data quality platform to test and assess and measure and profile the quality of your data with actual, you know, explicit data quality scores. So data quality is something we take very seriously, and I think we have some of the best data in the business. So I think that's probably the thing that's most important is no matter anything else, you know, the data vendor promises, if the data is not reliable, then you can't rely on it. And it's you know, you get yourself into a garbage in, garbage out situation.

Thank you, Mick. So the last question we have here. If anyone has any more questions, please do send them in now before we end the session. One more question.

In terms of what it takes to build, mentioned on-site, too, often non-technical leaders and executives are starting to ask if using AI coding has tipped the scale in favor of in house builds. What's your experience? And is there any data you can share, e.g. headcount change, cost, timeline?

Yeah. Good question. And we'd be remiss without talking about AI at some point. Right? It's like it enters every conversation these days.

You can't avoid it. Interesting question. Here's some of my thoughts.

Your team is certainly going to be empowered by using Gen AI tools like Callic Cloud Code. But remember, your vendors have access to the same tools. Right?

So the way I look at it is that everyone is leveling up. Right? So, yes, you can move much faster than a vendor could a year ago, but the vendor can move much faster as well. It does reduce the amount of time it takes to produce code, but code is just one step in the overall life cycle.

I think it places even more emphasis on domain experience when you make code easier. In this case, capital markets expertise, market data expertise. That's gonna become more and more the differentiator. OneTick has more than twenty years of experience; capital markets is all we do.

You know, we're not a vendor that builds generic data platforms or databases. We are a capital markets, market data focused vendor with deep, deep capital markets expertise.

So, you know, we're also aggressively adopting GenAI techniques at OneTick. And so, yes, it is easier than ever to build things, but that domain expertise and that industry experience is still highly valuable. It is not something that can be replicated by GenAI.

Mick, thank you so much for all the preparation that went into this event today. And thank you to all the people who registered and submitted questions, and participated in the polls. Now they have some food for thought, Mick, and thank you for this. We'll follow-up with some questions, some additional polls and surveys, and get feedback from everybody here.

If you have colleagues who are interested in this topic, please do share this recording. We're happy to answer their questions as they come up. 

Thought provoking. Thank you, Mick.

Thank you, everybody. Appreciate your time.

More About OneTick Cloud:

OneTick Cloud provides real-time, intra-day and historic data and analytics on trading activity leading to actionable insights for sales, trading, and surveillance. With hundreds of customers, you know that OneTick is a recognized platform you can trust. We will cover these many powerful advantages to OneTick, and show you how, together, we can meet your business needs.