JEALOUS ABOUT NIXTLA

Before we start, I just want to mention that this is a copy of the same article I wrote in my personal blog. We are re-starting Data Intelligence consulting work in Latin America, and I thought only proper to start talking about companies that inspire us and help us in our mission to bring Data Science and Business Analytics to companies to help them bridge the gap between Europe, Asia, and US.

I am a little jealous about a company I just discovered, called Nixtla. You can check their website here. According to their own words, “… Nixtla democratizes access to state-of-the-art predictive insights, eliminating the need for a dedicated team of machine learning engineers.”

Screenshot

Tall order and one I am a little doubtful about because I am so used to people with no or minimal statistical skills, basically butchering forecasting, but hey, maybe that’s just me, and I had no luck in the real world.

The one thing I am so envious of Nixtla is their love for open source and their large codebase contribution to Github. For a company that wishes to democratize predictive analytics, they sure put their money where their mouth is, because the open way they are giving away code and examples is enough for more than one consultant to open their forecasting venture just on their excellent libraries. I am counting six repositories with a plethora of predictive analytical code, including:

  • StatsForecast offers a collection of widely used univariate time series forecasting models, including automatic ARIMA, ETS, CES, and Theta modeling optimized for high performance using numba. It also includes a large battery of benchmarking models.
  • mlforecast is a framework to perform time series forecasting using machine learning models, with the option to scale to massive amounts of data using remote clusters.
  • NeuralForecast offers a large collection of neural forecasting models focusing on their performance, usability, and robustness. The models range from classic networks like RNN to the latest transformers:  MLP,  LSTM,  GRU, RNN, TCN,  DeepAR,  NBEATS,  NBEATSx,  NHITS,  DLinear, NLinear, TFT, Informer, AutoFormer, FedFormer, PatchTST, StemGNN, and TimesNet.
  • HierarchicalForecast offers a collection of reconciliation methods, including BottomUp, TopDown, MiddleOut, MinTrace, and ERM. And Probabilistic coherent predictions including Normality, Bootstrap, and PERMBU.
  • tsfeatures calculates various features from time series data. Python implementation of the R package tsfeatures.
  • Nixtla, Open Source Time Series Ecosystem, which is basically a compendium of all above, plus some new classes (as far as I can see.)

And yes, I just copied and pasted the whole thing, not because I am lazy, but rather because I feel a little overtaken by this treasure chest of forecasting and predictive goodness. In an age of more proprietary code and tools, such a degree of technical philanthropy is not unheard of, but not common either. As I said, I feel jealous of Nixtla for being a company so far advanced in predictive analytics that they can contribute to the wider world of data science and make a big impact on the rest of the community. But then, that is certainly something worth being jealous of.

Follow Nixtla on Twitter here.