Skip to main content
skill CX Codex

analyzing-time-series — Claude Code skill: time series diagnostics

analyzing-time-series runs the statistical checks on a CSV time series and tells you whether it is stationary, seasonal or worth differencing before you pick a forecasting model.

no license 1 installs ★ 1,438 updated 2026-06-08

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add https-deeplearning-ai/sc-agent-skills-files --skill analyzing-time-series -a claude-code
Open on GitHub

What analyzing-time-series does

Point it at a CSV time series and it runs the checks you would otherwise do by hand before fitting a forecast. The file needs two columns, one with dates and one with numbers. Names are detected automatically unless you pass --date-col and --value-col.

Start with python scripts/diagnose.py data.csv --output-dir results/. That leaves diagnostics.json with the metrics plus summary.txt written for a person to read. An optional second pass, python scripts/visualize.py data.csv --output-dir results/, writes plots into results/plots/: ACF/PACF, a decomposition, rolling statistics, box plots by weekday or month. Run it after the diagnostics, because the plots are synced to the stationarity results.

Interpretation guidance sits in references/interpretation.md and covers forecastability, needed differencing and seasonal period. The scripts import pandas, numpy, matplotlib, statsmodels and scipy, so install those first or nothing runs. --seasonal-period overrides the detected period.

What it helps with

  • Check stationarity before choosing a forecasting model
  • Find the seasonal period from the data itself
  • Decide how many differences the series needs
  • Generate ACF, PACF and decomposition plots in one run
  • Tell whether a series is forecastable at all

Example prompts for the agent

  • Run the time series diagnostics on sales.csv and tell me if it is stationary
  • Analyze data.csv before I fit a forecast, put the output in results/
  • Plot ACF and PACF for temperature.csv and say what seasonal period it shows

How to install analyzing-time-series

1Skills CLI (Vercel)

npx skills add https-deeplearning-ai/sc-agent-skills-files --skill analyzing-time-series -a claude-code

2Manually: put the files in place

~/.claude/skills/analyzing-time-series/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/analyzing-time-series/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about analyzing-time-series

How do I install analyzing-time-series in Claude Code?

Run npx skills add https-deeplearning-ai/sc-agent-skills-files --skill analyzing-time-series -a claude-code. The skill lands at ~/.claude/skills/analyzing-time-series/SKILL.md.

What does analyzing-time-series do?

It tests a CSV time series for stationarity, seasonality and trend, and judges how forecastable the series is. Results go to diagnostics.json, summary.txt and a folder of plots.

What does the skill need installed?

The scripts rely on pandas, numpy, matplotlib, statsmodels and scipy. Without those Python packages the diagnostic and plotting steps will not start.

Similar skills

All skills for Data and databases →

Project sponsors

Companies that keep this analytics open