# Arturus/kaggle-web-traffic

1st place solution

Repository: https://github.com/Arturus/kaggle-web-traffic
Canonical: https://ross.abutalabs.com/products/kaggle-web-traffic
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: kaggle-web-traffic, kaggle, time-series, timeseries, rnn-encoder-decoder, rnn, tensorflow, cudnn, cocob, seq2seq
Last push: 2022-10-09T07:10:28+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3211, "days_push": 1424, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1849, forks 657 (observed 2026-08-28T04:05:44.142789+00:00)

## What it is
The 1st place solution code for the Kaggle Web Traffic Time Series Forecasting competition, implementing an RNN encoder-decoder (seq2seq) model in TensorFlow. It includes feature building, training, and prediction pipelines, and serves as a reference for time-series forecasting with deep learning.

## Use cases
- forecast web traffic time series
- learn how to build a seq2seq time-series model
- reproduce a Kaggle competition winning solution
- study RNN encoder-decoder forecasting in TensorFlow
- apply deep learning to multi-series forecasting

## When to choose
- you want a proven reference architecture for time-series forecasting with RNNs
- you are competing in or studying the Kaggle web traffic forecasting problem
- you have GPU resources and want to reproduce the competition results

## When to avoid
- you need a production-ready forecasting library or maintained tool
- you lack a GPU, since training requires cuDNN
- you want a general-purpose time-series package rather than competition-specific code

## Facets
- artifact type: learning-resource
- maturity: maintenance
- function: machine-learning, deep-learning, llm-training
- domain: time-series, machine-learning, data-science
- platform: python
- tags: kaggle-competition, seq2seq, rnn-encoder-decoder, tensorflow, time-series-forecasting, cudnn, winning-solution, gpu, linux

## Member repositories
- Arturus/kaggle-web-traffic (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:44.142789+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:17:15.524495+00:00, confidence not recorded.
  - readme: https://github.com/Arturus/kaggle-web-traffic (fetched 2026-08-28T04:05:44.142789+00:00, sha 2376dad11aae)
- Data as of 2026-08-30T08:39:29.467469+00:00.