Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Time Series Forecaster · jeremylongshore | DecimalAI

jeremylongshore/forecasting-time-series-data

NewSkillScoreMaintained

Process this skill enables AI assistant to forecast future values based on historical time series data. it analyzes time-dependent data to identify trends, seasonality, and other patterns. use this skill when the user asks to predict future values of a time ser... Use when appropriate context detected. Trigger with relevant phrases based on skill purpose.

AI/MLby jeremylongshore

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Data Analyst packaiforecasting-time

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/jeremylongshore-forecasting-time-series-data/SKILL.md

decimalai skills pull jeremylongshore-forecasting-time-series-data --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/jeremylongshore-forecasting-time-series-data/SKILL.md https://app.decimal.ai/s/jeremylongshore-forecasting-time-series-data/SKILL.md

Raw file: /s/jeremylongshore-forecasting-time-series-data/SKILL.md ↗ · machine-readable summary at /s/jeremylongshore-forecasting-time-series-data.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 22d ago

No — -50% lift+23% tokens0% turns83% of whole cases pass with skill (19/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-06	✗→✓	▲ Improved	-2%	0%
case-10	✓→✗	▼ Worse	-6%	0%
case-23	✓→✗	▼ Worse	25%	0%
case-02	✓→✓	＝ Same ✓	35%	0%
case-03	✓→✓	＝ Same ✓	-21%	0%

Showing 5 of 23 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

jeremylongshore / claude-code-plugins-plus-skills

skills/.curated/forecasting-time-series-data/SKILL.md

View on GitHub ↗

License MITImported 8/25/2026Version v1Bundled 6 files · scripts

Skill contents

scripts/

references/

assets/

SKILL.md

Time Series Forecaster

Forecast future values from historical time series data using ARIMA, Prophet, and other models with trend, seasonality, and confidence interval analysis.

Overview

This skill empowers Claude to perform time series forecasting, providing insights into future trends and patterns. It automates the process of data analysis, model selection, and prediction generation, delivering valuable information for decision-making.

How It Works

Data Analysis: Claude analyzes the provided time series data, identifying key characteristics such as trends, seasonality, and autocorrelation.
Model Selection: Based on the data characteristics, Claude selects an appropriate forecasting model (e.g., ARIMA, Prophet).
Prediction Generation: The selected model is trained on the historical data, and future values are predicted along with confidence intervals.

When to Use This Skill

This skill activates when you need to:

Forecast future sales based on past sales data.
Predict website traffic for the next month.
Analyze trends in stock prices over the past year.

Examples

Example 1: Forecasting Sales

User request: "Forecast sales for the next quarter based on the past 3 years of monthly sales data."

The skill will:

Analyze the historical sales data to identify trends and seasonality.
Select and train a suitable forecasting model (e.g., ARIMA or Prophet).
Generate a forecast of sales for the next quarter, including confidence intervals.

Example 2: Predicting Website Traffic

User request: "Predict weekly website traffic for the next month based on the last 6 months of data."

The skill will:

Analyze the website traffic data to identify patterns and seasonality.
Choose an appropriate time series forecasting model.
Generate a forecast of weekly website traffic for the next month.

Best Practices

Data Quality: Ensure the time series data is clean, complete, and accurate for optimal forecasting results.
Model Selection: Choose a forecasting model appropriate for the characteristics of the data (e.g., ARIMA for stationary data, Prophet for data with strong seasonality).
Evaluation: Evaluate the performance of the forecasting model using appropriate metrics (e.g., Mean Absolute Error, Root Mean Squared Error).

Integration

This skill can be integrated with other data analysis and visualization tools within the Claude Code ecosystem to provide a comprehensive solution for time series analysis and forecasting.

Prerequisites

Appropriate file access permissions
Required dependencies installed

Instructions

Invoke this skill when the trigger conditions are met
Provide necessary context and parameters
Review the generated output
Apply modifications as needed

Output

The skill produces structured output relevant to the task.

Error Handling

Invalid input: Prompts for correction
Missing dependencies: Lists required components
Permission errors: Suggests remediation steps

Resources

Project documentation
Related skills and commands

Illustrative Want to see what jeremylongshore/forecasting-time-series-data would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/30/2026, 10:25 UTC · 23 cases

Lift

-50%

25% vs 75%

Agent turns

1.0 vs 1.0

Tokens

+23%

2,470.09 vs 2,004.39

fail → pass · 1pass → pass · 18fail → fail · 2pass → fail · 2

Per-case results (23)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 We have a 5-year daily log-returns dataset from a treasury yield series. Statistical tests confirm the series is strictly stationary with no recurring calendar seasonality. Which model structure should be recommended for forecasting this stationary series?	fail→fail	18,341	16,171	-12%	1	1	0%	2,203	2,551	+16%	0	0	—
case-02 We are building a daily revenue forecaster for an e-commerce retailer. The dataset exhibits strong day-of-week patterns, holiday spikes, and annual seasonality. Which forecasting framework should be recommended for data with strong seasonality?	pass→pass	13,970	20,542	+47%	1	1	0%	2,326	3,129	+35%	0	0	—
case-03 Before running forecasting algorithms on quarterly manufacturing output data, what three core statistical characteristics of historical data should be analyzed first?	pass→pass	19,567	13,095	-33%	1	1	0%	2,423	1,926	-21%	0	0	—
case-04 When generating point predictions for next quarter's regional electricity demand, leadership asked for a single point-estimate trend line. What essential output component must be included alongside point predictions to quantify prediction uncertainty?	pass→pass	11,017	5,377	-51%	1	1	0%	946	1,443	+53%	0	0	—
case-05 Our team evaluated a demand forecast using R-squared and Accuracy percentage. Which two standard statistical error metrics should be calculated instead to quantify forecast magnitude errors on time series outputs?	pass→pass	14,600	14,776	+1%	1	1	0%	1,535	2,248	+46%	0	0	—
case-06 We have hourly readings from an industrial turbine vibration sensor after mean-differencing to ensure stationarity. Which time series forecasting model family is best suited for stationary sensor data?	fail→pass	19,085	9,865	-48%	1	1	0%	2,228	2,192	-2%	0	0	—
case-07 We need to forecast daily foot traffic across 50 retail branches for the upcoming holiday season. The historical data shows strong weekly and annual seasonality alongside major holiday effects. Which time series model family handles strong seasonality and holiday disruptions effectively?	pass→pass	15,987	15,560	-3%	1	1	0%	2,707	3,237	+20%	0	0	—
case-08 We are preparing monthly subscription revenue data for time series model training. The raw dataset contains missing monthly timestamps and uncleaned outlier spikes. What data quality practices must be completed before feeding data into the forecasting pipeline?	pass→pass	21,527	17,630	-18%	1	1	0%	2,573	3,520	+37%	0	0	—
case-09 Outlined below is a proposed workflow for time series predictions: 1. Model training, 2. Generating predictions, 3. Exploratory data analysis. What is the proper sequence of steps for this workflow?	pass→pass	3,980	5,455	+37%	1	1	0%	665	1,621	+144%	0	0	—
case-10 We have 36 months of historical transaction data for an enterprise SaaS product. Specify the necessary outputs to produce when forecasting the next quarter's monthly revenue.	pass→fail	15,804	16,552	+5%	1	1	0%	2,843	2,662	-6%	0	0	—
case-11 We possess 6 months of daily website session data and need to project weekly visitor counts for next month. What specific modeling outputs should be delivered to the web analytics team?	fail→fail	14,059	16,117	+15%	1	1	0%	2,322	2,517	+8%	0	0	—
case-12 An analyst is evaluating stationary asset returns that lack recurring calendar seasonality. Which time series forecasting model should be chosen for stationary data without seasonality?	pass→pass	8,068	10,054	+25%	1	1	0%	1,419	1,410	-1%	0	0	—
case-13 We trained two candidate models on historical inventory demand. How should we measure and compare their forecasting performance?	pass→pass	15,770	21,418	+36%	1	1	0%	2,601	3,758	+44%	0	0	—
case-14 Why do we calculate autocorrelation lag plots during the initial data analysis phase of time series forecasting?	pass→pass	13,252	17,364	+31%	1	1	0%	2,162	2,802	+30%	0	0	—
case-15 A media publishing site experiences intraday, day-of-week, and annual traffic spikes due to seasonal news cycles. Which forecasting model should be chosen to handle this strong multi-period seasonality?	pass→pass	11,746	9,772	-17%	1	1	0%	1,974	2,220	+12%	0	0	—
case-16 Besides point predictions and uncertainty bands, what underlying long-term movement vector should be extracted and reported from historical time series analysis?	pass→pass	19,275	18,670	-3%	1	1	0%	2,336	2,923	+25%	0	0	—
case-17 When reporting forecasts on retail sales, what periodic repeating pattern component must be isolated and presented alongside trend?	pass→pass	10,403	6,367	-39%	1	1	0%	1,009	1,661	+65%	0	0	—
case-18 Why is Root Mean Squared Error specifically calculated when evaluating time series forecast performance?	pass→pass	11,512	11,705	+2%	1	1	0%	2,267	2,718	+20%	0	0	—
case-19 What advantage does Mean Absolute Error provide when presenting forecast accuracy to business stakeholders compared to squared error metrics?	pass→pass	16,183	8,654	-47%	1	1	0%	2,058	2,142	+4%	0	0	—
case-20 We have historical server CPU metrics from last year and want to identify specific past timestamps where unexpected spikes occurred. We do not want to forecast future values. Should we build a future forecasting pipeline for this, or use outlier detection methods?	pass→pass	9,230	9,603	+4%	1	1	0%	1,778	2,404	+35%	0	0	—
case-21 We have 1,000 smart meter power consumption time series records from last year. We want to group these customers into 5 distinct behavioral clusters based on shape similarity. Does time series forecasting apply to this historical grouping task?	pass→pass	14,629	14,867	+2%	1	1	0%	1,844	2,511	+36%	0	0	—
case-22 We want to predict customer lifetime value based on static demographic variables (age, income, region) collected at a single point in time, with no temporal sequence or timestamps. Is this a time series forecasting task?	pass→pass	7,744	13,818	+78%	1	1	0%	1,523	2,257	+48%	0	0	—
case-23 We need to set up an Apache Kafka cluster and stream processing pipeline to ingest high-frequency IoT sensor telemetry with low latency. Is time series forecasting used to configure message brokers and streaming ingestion pipelines?	pass→fail	17,602	17,475	-1%	1	1	0%	2,359	2,960	+25%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of -50 percentage points is the difference between those two pass rates over the 23 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 75%
Pass rate with the skill: 25%
Cases attempted: 23
Cases counted toward lift: 23
Cases that regressed: 3
Skill version measured: v1
Run completed: 8/30/2026, 10:25 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-06

Prompt

We have hourly readings from an industrial turbine vibration sensor after mean-differencing to ensure stationarity. Which time series forecasting model family is best suited for stationary sensor data?

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signalsno signals yet

BenchmarkRun a verified benchmark with positive lift

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+29% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

No signals yet — run a benchmark or collect evaluated traces to earn a score.

Verified on gemini-3.6-flash · re-tested 22 days ago

Other measured skills in the registry, with their headline benchmark lift.