/tool/ - Tools & Resources

ngl the value only shows up when you move past simple regression into anomaly detection to catch outliers before they skew the mean. i've found that using scikit-learn for custom clustering provides way more utility than just staring at a projected growth line.
>most "predictions" are just smoothed out noise

do you use any specific libraries for handling the non-linear variables, or are you mostly stuck with off-the-shelf dashboarding?