Who should use the Time Series Forecasting workflow?
Teams or solo builders working on data tasks who want a repeatable process instead of one-off tool experiments.
AI Workflow · Data
A concise workflow for generating and validating time series forecasts using demand preparation, model execution, and interactive visualization.
Deliverable outcome
A production-ready forecasting system with ongoing performance tracking.
30-90 minutes
Includes setup plus initial result generation
Free to start
You can swap tools by pricing and policy requirements
A production-ready forecasting system with ongoing performance tracking.
Use each step output as the input for the next stage
Step map
Instead of relying on a single generic AI model, this pipeline connects specialized tools to maximize quality. First, you'll use Dataiku to a clean, regularly-spaced time series ready for analysis. Then, you pass the output to Dataiku to understanding of data patterns and a stationary series (or transformation plan) for modeling. Then, you pass the output to Tecton to a feature-rich dataset with a clear temporal train/test split. Then, you pass the output to Darts to trained models ready for out-of-sample evaluation. Then, you pass the output to Darts to quantitative performance metrics for each model on unseen data. Then, you pass the output to InfluxDB to a validated, interactive visualization that communicates forecast accuracy and business insights. Finally, Dataiku is used to a production-ready forecasting system with ongoing performance tracking.
Data Acquisition and Cleaning
A clean, regularly-spaced time series ready for analysis.
Exploratory Data Analysis and Stationarity Check
Understanding of data patterns and a stationary series (or transformation plan) for modeling.
Feature Engineering and Lag Creation
A feature-rich dataset with a clear temporal train/test split.
Model Selection and Training
Trained models ready for out-of-sample evaluation.
Forecast Generation and Backtesting
Quantitative performance metrics for each model on unseen data.
Result Validation and Interactive Visualization
A validated, interactive visualization that communicates forecast accuracy and business insights.
Deployment and Monitoring (optional)
A production-ready forecasting system with ongoing performance tracking.
Collect raw time series data from internal databases, APIs, or flat files. Handle missing values, outliers, and time zone inconsistencies to ensure a clean, continuous timestamp index.
Why Dataiku: Dataiku provides integrated data wrangling and cleaning capabilities, which directly match the needs of this step for Python/SQL-based data acquisition and cleaning.
Visualize the series to identify trends, seasonality, and anomalies. Perform statistical tests (ADF, KPSS) to check stationarity, and apply differencing or transformations if needed.
Why Dataiku: Dataiku includes data wrangling and cleaning tools that support exploratory data analysis, and its AutoML pipeline can assist with stationarity checks.
Create lagged variables, rolling statistics, and calendar features (day of week, month, holiday flags) to capture temporal dependencies. Split data into training and testing sets chronologically.
Why Tecton: Tecton is specifically designed for feature engineering for ML, including lag creation and feature serving for time series.
Choose one or more forecasting models (e.g., ARIMA, Prophet, LSTM, XGBoost) based on data characteristics. Train on the training set, tuning hyperparameters via time series cross-validation.
Why Darts: Darts is a comprehensive time series forecasting library that supports model selection and training with various algorithms.
Generate predictions on the test set (or future horizon) using each trained model. Compute forecast errors (MAE, RMSE, MAPE) and compare against a naive baseline.
Why Darts: Darts includes built-in backtesting functionality, which directly supports forecast generation and evaluation.
Plot actual vs. predicted values with confidence intervals (if available). Create an interactive dashboard (e.g., Plotly, Streamlit) to allow stakeholders to explore forecasts, residuals, and scenario adjustments.
Why InfluxDB: InfluxDB provides data visualization and monitoring, which can be used for interactive visualization of forecasts.
Package the best model into an API or scheduled job for ongoing forecasts. Set up monitoring to detect drift in data or model performance over time.
Why Dataiku: Dataiku supports model deployment and monitoring, aligning with the needs for Docker, FastAPI, MLflow, and Airflow integration.
§ Before you start
Teams or solo builders working on data tasks who want a repeatable process instead of one-off tool experiments.
No. Start with the top pick for each step, then replace tools only if they do not fit your pricing, compliance, or output needs.
Open the mapped task page and compare top options side by side. Prioritize output quality, integration fit, and predictable cost before scaling.
§ Related
Track competitor moves and market shifts in real-time with automated intelligence gathering — so you always know what your rivals are doing.
Connect siloed business applications into a unified, AI-managed operational pipeline that eliminates manual handoffs between systems.
Analyze portfolios, backtest investment strategies, and receive AI-generated market signals — giving individual investors access to institutional-grade tools.