Alternative Data

Trading Systems with Weather Data

Build trading strategies using 30 years of historical weather data — temperature, precipitation, HDD/CDD, agricultural zones, and PADD regions. The same alternative data edge that powers the world’s top commodity desks, now built into Build Alpha.

30+
Years of Data
19
Locations
$30B+
Citadel Commodity P&L

The Opportunity

Weather Data for Commodity Trading

Trading commodity markets like agriculture, energy, metals and even stock index futures can greatly benefit from additional non-price based contextual data. There has been a recent surge in professional trading operations acquiring weather data and weather patterns to predict and model commodity prices.

According to Bloomberg, some hedge funds are paying up to $1 million for weather modelers.

Bloomberg headline showing hedge funds paying up to $1 million for weather modeling experts

Case Study

How Citadel Uses Weather to Dominate Commodities

Citadel — the world’s most successful hedge fund — generates more from commodities than any other strategy: $30B+ since 2002. However, half of that gain has come in the last four years under Sebastian Barrack.

Barrack grew a tiny strategy to one that generates 50%+ of Citadel’s profit some years by using weather models and physical tracking. The Financial Times has a great piece on this: How Citadel harnessed the weather to claim hedge fund crown.

Build Alpha received a request to add weather data from a client who specializes in risk management in energy markets — which led to the development of the full weather signal library now included in Build Alpha.

Citadel commodities performance driven by weather data models under Sebastian Barrack

Deep Dive

The Complete Guide to Weather Data in Trading

Everything you need to know about weather signals, how they work, and how Build Alpha puts them to use. Updated .

Types of Weather Data for Trading Strategies

Build Alpha now contains tons of built-in weather signals that all update daily. No need to connect to a weather API or collect weather conditions yourself. Here is the weather data included in this first release.

Temperature Data

Raw Fahrenheit temperature data is useful, but comparing today’s temperature to the previous ten years’ worth of data to establish a baseline and better spot anomalies is wiser. Build Alpha also smooths the data so we are not just comparing May 6th to all previous May 6ths, but to the dates around May 6th ±10 days.

Reading Temperature Percentile

A simple reading of 88 would mean today’s May 6th temperature is in the 88th percentile of all closely related days over the past 10 years.

Percentile = rank(today’s temp, all ±10 day temps over 10 years) × 100

This normalization makes the data comparable across seasons, years, and locations — exactly what you need for systematic signal generation.

Precipitation Data

Precipitation can have a direct impact on agricultural yields. Too much and planting season can be delayed. Too little and crops become stressed.

Precipitation data can also lend clues into port congestion, cost changes to waterway transportation and even infrastructure risk, as mining or drilling operations can be paused due to washed-out roads in remote areas.

Behavioral finance may also be interested in rainfall and precipitation data. Does an overly wet season impact consumer spending? Does it force more energy demand as we stay in our homes? Can we make more informed trading decisions knowing how often it rained on Wall Street? Does weather affect investor mood?

Heating Degree Days and Cooling Degree Days

The energy industry has created Heating Degree Days (HDD) and Cooling Degree Days (CDD) to quantify heating and cooling fuel demand. Build Alpha has included both of these as well as a percentile rank for each.

Heating Degree Days measures how “cold” a day was relative to 65 °F and thus how much heating a home or business likely required.

Cooling Degree Days measures how “hot” a day was relative to 65 °F and correlates to the air-conditioning load.

Calculate Heating Degree Days (HDD)

HDD is defined as the maximum of zero or the difference of 65 degrees less today’s midpoint temperature.

Midpoint = (High + Low) / 2 = (40 + 20) / 2 = 30
HDD = max(0, 65 − 30) = 35 heating degree days

Calculate Cooling Degree Days (CDD)

CDD is defined as the maximum of zero or the difference of today’s midpoint temperature less 65 degrees.

Midpoint = (High + Low) / 2 = (95 + 65) / 2 = 80
CDD = max(0, 80 − 65) = 15 cooling degree days

Historical Weather Data Locations

The next step was determining specific locations to sample weather data from to derive actionable insights. Build Alpha’s vast network of traders helped inform these selections.

Major Cities

Build Alpha pulls data from a handful of major cities — each chosen for proximity to key trading floors, population centers, and energy infrastructure:

🏙️

New York City

Wall Street & financial hub

🌬️

Chicago

CME Group & derivatives hub

🛢️

Houston

Energy capital of the U.S.

☀️

Los Angeles

West coast population center

🌁

San Francisco

Tech & Pacific coast trade

🏛️

Washington DC

Policy & regulatory capital

Map of major U.S. cities for weather data in Build Alpha — New York, Chicago, Houston, Los Angeles, San Francisco, Washington DC

Major U.S. cities sampled for weather data in Build Alpha.

Agricultural Zones

Build Alpha also has weather data for six agricultural zones critical to crop production and commodity pricing:

🌽

Corn Belt

Iowa, Illinois, Indiana, Ohio

🌾

Delta States

Mississippi, Arkansas, Louisiana

🌿

Great Plains

Kansas, Nebraska, the Dakotas

🌲

Pacific Northwest

Washington, Oregon

🍑

Southeast

Georgia, Carolinas, Virginia

🍁

Northeast

New England & Mid-Atlantic

Map of U.S. agricultural zones approximated for Build Alpha weather signals — Corn Belt, Delta States, Great Plains, Pacific Northwest, Southeast, Northeast

U.S. agricultural zones used for weather-based commodity trading signals.

PADD Regions for Energy Trading

Finally, Build Alpha has weather data for the Petroleum Administration for Defense Districts (PADDs) determined by the U.S. Energy Information Administration (EIA). These regions are the standard geographic divisions used for tracking petroleum supply and demand across the United States.

PADD 1A

New England

PADD 1B

Central Atlantic

PADD 1C

Lower Atlantic

PADD 2

Midwest

PADD 3

Gulf Coast

PADD 4

Rocky Mountain

PADD 5

West Coast

Map of Petroleum Administration for Defense Districts (PADD) regions via EIA for energy trading weather data

Petroleum Administration for Defense Districts (PADDs) via the U.S. EIA.

Build Alpha Trading Strategy Using Weather Data

Often traders rely on market volatility, mistakes by other market participants, and an ability to manage risk to outperform. However, many system traders fail because their strategies do not have enough context regarding market regime or the current trading environment.

Adding context with non-price-based data is often a great way to improve strategy stability. This has been one of the greatest learnings in my early career — and as evidenced above by Citadel and others — and an incredible advantage of Build Alpha.

Here is a simple trading system with weather data on Corn Futures. It holds for one day and uses a volatility-based sizing method. It has a -0.02 correlation to S&P 500 buy and hold (correlation uses negative days and drawdowns only). These weather strategies can be a nice complement to an equities portfolio when the latter takes occasional significant losses.

Corn futures commodity trading strategy equity curve using weather signals in Build Alpha with volatility-based sizing

Corn Futures weather-based strategy in Build Alpha: 1-day hold, volatility-based sizing, -0.02 correlation to S&P 500.

Can Weather-Based Trading Strategies Be Automated?

Yes. Build Alpha can generate fully automate-able code for any strategy created. This generated code accesses updated weather data — and any other price and non-price-based data Build Alpha offers — in real-time. You simply select a strategy, hit the generate code button, and copy and paste it into your platform.

Trading platform logos showing Build Alpha code export support for TradeStation, MultiCharts, NinjaTrader, MetaTrader, ProRealTime, TradingView, and Python

Build Alpha exports fully automated code for all major trading platforms.

Create Your Own Weather Signals

Build Alpha also has a custom indicator editor (no code) with access to all this data so you can create and add any signal to the Build Alpha signal list with ease. The BA strategy engine can combine built-in and custom-made signals with various money management techniques, stop loss levels, position sizing methods, technical filters, and more.

The no-code editor can access weather events, market volatility measures, math functions, technical indicators, and more to create your own weather signals. For instance, you could sum HDD over the past N days to build a cumulative energy demand indicator.

You can also extend Build Alpha with Python and add any machine learning or custom signal you may already have.

Future Ideas for Build Alpha Weather Signals

Future development around weather-based signals could include extreme weather events, advanced climate models, weather forecasts vs. weather experienced, cloud cover data, solar power generation data, energy supply correlations to long-term weather forecasts, seasonal weather changes, and more.

Combining trading systems with weather data and Commitment of Traders Report data (and other non-price-based data) has been a ton of fun lately. I strongly recommend playing around with that! It is almost like the big players in the COT report are… well… looking at weather data…

I am excited to continue adding to Build Alpha’s historical weather database. If you have suggestions, improvements or additions please feel free to share! That is what makes the Build Alpha community so great and half the reason I do this.

Thanks for reading,
David

Updated .

Signals

Weather Signals in Build Alpha

All signals update daily with 30+ years of historical data. No API connections or data collection required.

Temperature Percentile

Smoothed 10-year percentile rank comparing today’s temperature to historical norms ±10 days. Spots anomalies across any season.

Precipitation

Daily precipitation data impacting agricultural yields, port congestion, waterway costs, infrastructure risk, and investor behavior.

Heating Degree Days

HDD = max(0, 65 − midpoint). Quantifies daily heating fuel demand. Includes raw values and percentile rank.

Cooling Degree Days

CDD = max(0, midpoint − 65). Measures air-conditioning load and energy demand. Includes raw values and percentile rank.

Agricultural Zones

Six crop-producing regions: Corn Belt, Delta States, Great Plains, Pacific Northwest, Southeast, and Northeast.

PADD Regions

Seven Petroleum Administration for Defense Districts for energy market analysis — the standard EIA geographic divisions.

FAQ

Frequently Asked Questions

What is weather data in trading?

Weather data in trading refers to using meteorological information — temperature, precipitation, heating/cooling degree days, and climate patterns — as alternative data signals to predict commodity prices, energy demand, agricultural yields, and even equity market behavior.

How does Citadel use weather data for trading?

Citadel has generated over $30B from commodities since 2002. Under Sebastian Barrack, a small weather-driven strategy grew to generate 50%+ of the firm’s profit in some years by using weather models and physical tracking. Read more: FT: How Citadel harnessed the weather.

What are Heating Degree Days (HDD) and Cooling Degree Days (CDD)?

HDD measures how cold a day was relative to 65°F and estimates heating fuel demand. CDD measures how hot a day was relative to 65°F and correlates to air-conditioning load. Both are standard energy industry metrics available as signals in Build Alpha.

What locations does Build Alpha cover?

Build Alpha covers 6 major U.S. cities (New York, Chicago, Houston, Los Angeles, San Francisco, Washington DC), 6 agricultural zones (Corn Belt, Delta States, Great Plains, Pacific Northwest, Southeast, Northeast), and 7 PADD regions for energy trading.

Can weather-based trading strategies be automated?

Yes. Build Alpha generates fully automated code that accesses updated weather data in real-time. Export to TradeStation, MultiCharts, NinjaTrader, MetaTrader 4/5, ProRealTime, TradingView, or Python for Interactive Brokers.

Can I create custom weather signals?

Yes. Build Alpha has a no-code custom indicator editor with access to all weather data, math functions, and technical indicators. You can also extend Build Alpha with Python to add machine learning or custom signals.

Why use non-price-based data in trading systems?

Many system traders fail because their strategies lack context about market regime and environment. Non-price-based data like weather, COT reports, and market breadth add contextual signals that improve strategy stability and reduce correlation to traditional equity portfolios.

How much weather history does Build Alpha include?

Build Alpha includes 30+ years of historical weather data across all locations and signal types. All data updates daily with no API connections or manual data collection required.

Ready to BUILD ALPHA?

Trade with the same weather edge powering the world’s top commodity desks.

No coding needed

Stop Wasting Time
Coding, Testing, Coding, Testing…

Build Alpha’s Genetic Algorithm uses your inputs to create 1000s of trading strategies in seconds

View Demo
Recent Posts
Get updates & ideas
Learn more about BA & get exclusive insights