The power of The Timeseries Refinery
All your time series, in one shared data referential. Accessible from a unified API point that transparently links all your data silos. Calculated series are all exposed in the catalog just like any other series.
Data storage is highly optimized for versioned data with simple time travel and backtesting capabilities. Access on the fly to external or legacy time series storage, without the need for data duplication.
A low code formula system to transform your raw data into enriched indicators. Time travel capabilities are enabled in all levels of aggregation. Extendable to your specific needs. The little streams make the big rivers.
When your data lake becomes a data swamp, choose to build a data mesh with the Timeseries Refinery. Specialize your instances around business oriented teams and aggregate their work in a federation system. Build a meta-catalog fed by various teams in a horizontal integration.
Build dashboards with simple configuration files. Choose between a low-level library to build complex and informative plots in a business-oriented fashion integrate in your own dashboard or a high-level library to build stand alone websites of plots.
Run any python and R code with machine learning algorithms. Schedule it on a large scale. Insert your fully-versioned forecasts, visualize them on the fly and make them available to your collaborators. Easy access to logs and launchers.
Build confidence in your data: audit the sources of any complex series. Explore the time series referential with multiple views ready to use. Look for any value of any revision of any series. In case of calculated series, explore recursively all the underlying components. Superpose multiple series on a whim to find correlations and covariations. Sometimes human expertise is the way to go : correct the data on the fly.
Evaluate the quality of your model with two dedicated API points: staircase and block-staircase. The required parameters are respectively a timedelta or a set of parameters with magic hours such as market opening/closing hour. Use the web UI for interactive backtesting of a collection of models and produce reports at scheduled time.
Manage all your processes in one single app. Give your analysts autonomy to supervise the IT processes. Launch and schedule any python code (from scraper to models) and access the logs easily. Easily identify the failing critical tasks. Configure the parallelisation or serialization of your tasks
No matter who you are
Organize the chaos of available data. Explore and transform your data with simple and efficient tools. Automate your data process in one single place. Extract values from the data by building high level indicators. Manage and share your data referential. Enjoy the data referential built by other specialists.
Explore new business opportunities faster and develop better performance for your teams. Share comprehensive, high-level data with your management to help them drive their businesses in an ever-changing environment. With 100% data traceability you are always in the driving seat.
Enjoy a ready to use tool, without new schema or API definition. Let the user be autonomous from data transformation to process management. Exploit the high modularity of the code (and its openness) to fine tune The Timeseries Refinery to your organization needs.
Our offer
Take a fresh start with a powerful tool
We can help you with the installation of the Timeseries Refinery in your own infrastructure. We can also host your Timeseries Refinery on our own cloud.
Learn to integrate the Timeseries Refinery within your teams: documentation is a start but good practices will launch you to the stars.
We can build connectors to your own database, data feeds, web subscriptions and private API’s. We can also develop specific functionalities like custom access rights or specific interface, etc.
Depending on your choice of deployment, we can provide full maintenance of your timeseries refinery or assist your IT to manage the tool.
Integrate easily ready to use time series from our catalog
CO2 emission forecast from power consumption designed to correctly stimulate and value demand response in 12 countries in EU. For more informations, go to eFlower website.
5 days ahead forecast of EEX Day Ahead power market prices
Country level wind and solar production for 12 countries in EU
ENTSO-E, RTE, ECMWF, GFS, ERA-5…
Open source
Built for developers
Homogeneous through the whole catalog, independently of the nature of the series.
API docDedicated to versioned timeseries. Highly optimized for such operations, from storage to http transfer, accessible from the tip of your fingers: get any version equally as fast.
All data management operations accessible through the web UI are also launchable through the python API: data creation, formula definition and evaluation, and task management.
Behind The Timeseries Refinery
Pythonian is a cross disciplinary team with various area of expertise : Software engineering, Data Management, Modelisation, Energy markets, Meteo/Hydro
Designed and used by experts from: