log.qmd

---
title: "Log"
description: "Process log"
sidebar: false
code-block-bg: false
code-block-border-left: true
format: 
  html:
    toc: true
    toc-depth: 3
    code-fold: false
    code-tools: false
    highlight-style: gruvbox
---

:::{.callout-note}
## Priority rank

1. Time series: classical method and ML
2. Time series: scaling up
3. ML on Cloud
4. Other ML methods such as classification 
5. A/B testing
:::


### Time series 

|Stage |Tasks |Progress | Python | R|
|--|---|----|-|-|
|Set up | Find a problem set |Done. FAO TS questions|||
|  |Find a dataset to test | Done. Kaggle rainfall data, with 4 variables to model|||
|  ||Advanced: Norsyss data, weekly with different locations and metrics |||
|Analysis: rainfall |EDA and processing |Done | Yes|Yes |
| |Model: classical |ARIMA <br/> SARIMA <br/> Exponential smoothing | | |
| |Model: ML | LSTM <br/> xgboost | | *keras in R is creating issues, ignore*|
|Analysis: norsyss | | | | |