DataScapR: A toolbox for stock market sonification
Hello
My name is Samuel Van Ransbeeck. I am doing my PhD on sonification art and I have recently competed a toolbox to use stock market data in composition. As it uses Bach a lot, I thought I'd share it here with you. I think it can be interesting. You can download the toolbox on https://datascapr.wordpress.com
If you want to use it (and later fill in the survey) that would be great!
What is DataScapR:
DataScapR is a MaxMSP toolbox for stock market sonification. The toolbox allows you to use historical and real-time data to create musical material or drive musical processes. While DataScapR offers an out-of-the-box operation, the modular nature makes it very easy to extend the patches as you wish. DataScapR is available for download and comes with a comprehensive manual as well as documentation within the patchers.
Great work, Samuel!
Thanks for sharing!
I made a little bit similar project (signification of stock data in realtime + simple motion tracking as an interface) some time ago. If you are interested: http://paweljanicki.jp/projects_aleastock_en.html
For those of us interested in your work, but not MaxMSP users: Is there anywhere that we can hear example sounds or view example movies of your UI?
Thanks!
Not yet, I will be working on that in the coming weeks.
Thanks! Let me know.
I'd be interested to know how your approach differs from Ciardi's (ICAD 2004). I've had a long interest in sonification of data
Ciardi is a funny story. Years ago, when I was getting started with sonification, I asked him if he had more information about the software or even some music. His answer was that he wanted to try to sell the software for use in the real market. Hence I developed StockWatch and now DataScapR. From his paper, it seems that our software has some similarities however, since 20034 Max has evolved a lot and thus offers more possibilities. With DataScapR you can use real-time data or historic data, control a VST/MIDI synth or create symbolic scores. Things like that are not mentioned in sMax. Nevertheless, sMax was an interesting thing, a pity that it did not get the attention it should have had.
On the update time: Ciardi speaks of 200-600 milliseconds updates. I do not know where he gets his data (I get mine from Yahoo Finance) but having tested my system, I found out that 30 seconds seemed to be the minimum time. Anything less than that would not yield meaningful changes in the data. In very high trading stocks you can get quite some variation but in more normal stocks, the price stays stable. Of course for historical data this problem does not arise, however, when using monthly data instead of daily data, you will get bigger differences between the data points.
Interesting. Are you trying to understand the market behavior through sound and if so, how many and what ticker parameters do you use and and what mapping choices did you make for sound generation?
The goal of this toolbox is to let composers use market data in an easy way to drive musical processes. It is not intended for trading on the markets for real (although I will expire that in the future). You should see DataScapR as giving a perspective on what happens in the markets. It can give you unexpected musical results, that you would not get with other compositional approaches. I think Steve Roden 's (a sound artist) quote conveys what I want to do: "Just as the installation would look and sound different if the data source was the ocean or Shakespeare; ear(th) in no way attempts to illustrate the earth's movement or to recreate an earthquake experience through sound. I am much more interested in simply allowing the earthquake data to generate a sound composition and to allow for my own misreadings of the data to suggest placements, sound ideas, performances, and sculptural forms. For me, this process is a kind of alchemy - to allow the materials to be transformed into something completely connected to, yet seemingly distant from, the source"
Mapping parameters: in the real-time data you can get 53 data points per stock, in the historical data you get 8 data points per stock (opening price, closing price, ...).
I tried to make the software as neutral as possible (hence the absence of examples) to avoid that I would be driving the composer in a certain direction. Later on I will publish works made using DataScapR but in this phase I think it is best to let the composer explore without being hindered by directional examples.
I just updated DataScapR, I made some adjustments in the MIDI mapping of realtime data. You can get the new version via datascapr.wordpress.com
Sorry for the errors, it was an oversight.
Very interesting --- I mis-asked (is that a word?) my question about mappings. I didn't mean to ask what parameters are available for mapping. I was interested to know to what audio-related attributes were you mapping available input parameters. When I worked on this stuff in the early 90s (and my focus was using audio to understand data rather than to create music) a couple of things became clear very quickly.
1) It was much better to keep the pitch constant and change other parameters. If you tied pitches to different tickers, you could follow and compare changes to multiple tickers simultaneously (much like two opera singers can sing different words at the same time and you can follow both)
2) Second order modulations turned out to be very powerful both on inputs and outputs. So for example you might have a vibrato attached to a particular pitch and change its rate based not on the value of some incoming parameter but rather based on how often that incoming parameter arrived, or how often it changed or on how often the difference between two successive values changed. Sometimes it was even better to change the rate at which the rate was changing!
I'm hoping I can find some time to play with your toolbox - thanks for making it available.
I have released an update for DataScapR which improves quite a lot of things. You can get it on the blog: datascapr.wordpress.com
I also made a video of an installation I am doing: Vapourwaves which uses eight stocks to control oscillator banks, a granularisation and a Disklavier. You can see it on the blog as well.
Did. you ever find a new way to access the data? or a new version for this? I would love realtime market data into max/msp. @THINKSAMUEL
https://polygon.io/?fbclid=IwAR1hMZ8iqlJURgZKCdOT7JyTEt-bII2bXr8jqlbwbEvMjDGrwpbkqJQSuzo
you could use node for this. Scroll down to the Node For Max section for a thorough overview - https://cycling74.com/articles/javascript-a-resource-guide-for-max-users
@axiomcrux supporting what Rob wrote, this project is similar and used Node:
https://cycling74.com/forums/binance-crypto-exchange-in-max-via-node-for-max-work-in-progress