All this code is very experimental, not even close to having a release, has flaws all over. WIP
Different ways to get support and resistance (horizontal) price zones in market data dataframes.
In this approach we are dealing with little information about 1000 tikers. At first I do not expect good results from this but it is interesting to try.
Currently there is one OHCL approach with a sklearn lib MeanShift clustering which is fast and it could be fine tunned to obtain better results.
This are current plots, which are not very usefull. Based on not very tunned settings, only 999 tickers of data.
