I'm not sure what the appropriate terminology is, but when trying to run simulations, it is always difficult for me to create good fake data.
I don’t have a specific application for this, but let me say that I want to play with some kind of stupid stock market forecasting algorithm - if I just used a standard random number generator to get my test data, it's all moving around .5, even through short periods of time, and this would not lead to the actual form of data that the stock market usually produces during the day (comparing it with stock charts). Even if the market closes without profit or loss, you can still find volatility in the middle - simple random walks do not create the same effects.
I think you could add rngs on top of each other, a larger value for a full day value, a smaller value per hour and a value still per second, summing them all together to get a more step-like pattern, but it's really too predictable - you know, as a developer, where these steps will be, or are likely to be, if you randomize the duration.
You could literally imitate the individual identities of the buyer and seller, I think, but this is a lot of work and calculations. (As far as I know, real stock market data is not available in raw form)
So, can we find free, easily accessible, fast, "interesting" data?
source
share