Matlab Access Datafeed

MathWorksOfficialCreated by MathWorks

Skill
Tool integration
Data and AI
BSD-2-Clause
No issues found

From the author

Connect to Bloomberg, FRED, Haver Analytics, LSEG Datastream, NOAA Climate Data Online, or US government agencies (BEA, BLS, Census, EIA) from MATLAB and retrieve data. Covers market data (bloomberg/blp/bloombergHypermedia), economic data (fredrs, haver/haverdirect/haverview), historical financial data (datastreamws), weather/climate data (@noaa from GitHub), and government data (usbea/usbls/uscensus/useia from GitHub).

— the description by MathWorks

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

The Datafeed Toolbox provides a unified interface for MATLAB users to retrieve data from various external providers, including Bloomberg, FRED, Haver Analytics, LSEG Datastream, NOAA, and US government agencies like the BEA, BLS, Census, and EIA. This skill covers the necessary connection patterns, interface selection, and data retrieval methods for these services, enabling seamless integration of external time-series and market data into MATLAB workflows.

What it does

Connects to Bloomberg for market and reference data using C++ or REST APIs.
Retrieves economic time series from the St. Louis Fed (FRED) via the fredrs interface.
Fetches global economic and financial data from Haver Analytics.
Accesses historical financial data from LSEG Datastream.
Retrieves historical climate and weather data from NOAA Climate Data Online.
Provides interfaces for US government agency data including BEA, BLS, Census, and EIA.

When to use it

When you need to integrate real-time or historical market data into your MATLAB analysis.
When performing economic research requiring time series from FRED or Haver Analytics.
When working with historical financial data from LSEG Datastream.
When your research requires climate or weather data from NOAA.
When you need to pull public economic indicators from US government agencies.

Advantages

Unified access to diverse data sources within the MATLAB environment.
Supports both local terminal-based connections and platform-independent cloud/REST APIs.
Provides specific interfaces optimized for different license types and environments.
Enables programmatic data retrieval for automated research and modeling pipelines.

Usage examples

“How do I connect to Bloomberg using the recommended C++ interface?”: Use the bloomberg constructor to initialize the connection, then call getdata to retrieve security information.
“How can I fetch historical data from LSEG Datastream?”: Initialize a datastreamws connection object with your credentials and use the history method to retrieve time series data.
“What is the correct way to get FRED data in MATLAB?”: Use the fredrs interface, which is the current standard for accessing FRED economic data in MATLAB R2026a and later.
“How do I access US government data like BLS employment statistics?”: Use the usbls interface from the GitHub-based collection to query specific employment data series.

Install

Download the skill as a ZIP:
https://autogineer.com/agent-skills/matlab/matlab-access-datafeed/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip matlab-access-datafeed.zip -d .agents/skills

Documentation by MathWorks

Datafeed Toolbox — Bloomberg, FRED, Haver Analytics, LSEG Datastream, NOAA, and US Government Data

Access financial, economic, climate, and government data from multiple providers through the MATLAB Datafeed Toolbox and related GitHub interfaces.

When to Use

User wants to connect to Bloomberg, FRED, Haver Analytics, LSEG Datastream, NOAA, or US government agencies (BEA, BLS, Census, EIA) from MATLAB
User needs to retrieve market data (prices, history, real-time, reference data) via Bloomberg
User needs to retrieve economic time series from the St. Louis Fed (FRED)
User needs to fetch economic/financial time series from Haver Analytics databases
User needs to retrieve historical financial data from LSEG Datastream
User needs help choosing between interfaces within a provider (e.g., bloomberg vs blp vs bloombergHypermedia, haver vs haverview)
User needs to retrieve historical climate or weather data from NOAA Climate Data Online
User needs to retrieve GDP, employment, population, or energy data from US government agencies
User is writing MATLAB code involving securities, economic indicators, climate data, government data, or data provider connectivity

When NOT to Use

General financial modeling or portfolio optimization that doesn't involve data provider connectivity
Bloomberg Terminal keyboard shortcuts or Excel add-in questions (not MATLAB)
Datastream Excel add-in (not MATLAB)
Real-time streaming from LSEG (datastreamws is historical only)
NOAA forecast API (weather.gov) or NOAA CO-OPS tidal data (different services from CDO)
Direct REST/HTTP calls to US government agency APIs without MATLAB
Agency-specific data interpretation (what BLS series IDs mean, Census variable codes, etc.)
Data sources not covered here (e.g., Yahoo Finance)
Installing MATLAB, Datafeed Toolbox, or toolbox products → use matlab-install-products
Querying relational databases via SQL, JDBC, or ODBC → use matlab-use-database
Debugging generic MATLAB runtime errors not specific to a data provider → use matlab-debug-code

Provider Overview

Provider	Data Type	Interfaces	Reference
Bloomberg	Market data (equities, fixed income, derivatives, indices)	`bloomberg`, `blp`, `bloombergHypermedia`, `bloombergEMSX`	`references/bloomberg.md`
FRED	U.S. economic data (GDP, unemployment, CPI, interest rates)	`fredrs` (replaces obsolete `fred`)	`references/fred.md`
Haver Analytics	Global economic & financial data (macro, industry, surveys)	`haver`, `haverdirect`, `haverview`	`references/haver.md`
LSEG Datastream	Historical financial data (equities, indices, economics, expressions)	`datastreamws`	`references/datastreamws.md`
NOAA CDO	Historical climate/weather data (temperature, precipitation, snowfall)	`noaa` (GitHub)	`references/noaa.md`
US Government Data	GDP, employment, population, energy (BEA, BLS, Census, EIA)	`usbea`, `usbls`, `uscensus`, `useia` (GitHub)	`references/us-govt-data.md`

Quick Interface Selection

Bloomberg — see `references/bloomberg.md` for full API patterns

Windows Desktop/Terminal user → bloomberg (C++)
Need platform independence (Data License) → bloombergHypermedia
Server/cloud, no Terminal (BPIPE) → bloombergBPIPE
Existing blp users → migrate to bloomberg (same API, no Java needed)

FRED — see `references/fred.md` for full API patterns

Any platform → fredrs (requires R2026a+, API key)
fred is obsolete — do NOT use it

Haver Analytics — see `references/haver.md` for full API patterns

Windows with local data files → haver
Windows needing remote access → haverdirect
Non-Windows / server / cloud → haverview (REST API)

LSEG Datastream — see `references/datastreamws.md` for full API patterns

Any platform → datastreamws (REST API, requires Datastream credentials)

NOAA Climate Data Online — see `references/noaa.md` for full API patterns

Any platform → noaa (GitHub interface, requires free CDO API token)
Clone from github.com/mathworks/noaa and addpath — not part of Datafeed Toolbox

US Government Data — see `references/us-govt-data.md` for full API patterns

BEA → usbea (GDP, national/regional/international economic accounts)
BLS → usbls (employment, unemployment, CPI, wages)
Census Bureau → uscensus (population, ACS, economic census)
EIA → useia (electricity, petroleum, natural gas, coal)
Clone from github.com/mathworks/Access-Public-US-Government-Data-from-MATLAB and addpath — not part of Datafeed Toolbox

How to Use References

Before generating code for a specific provider, read the corresponding reference file for detailed API patterns, function signatures, gotchas, and examples:

Working with Bloomberg data → read references/bloomberg.md
Working with FRED economic data → read references/fred.md
Working with Haver Analytics data → read references/haver.md
Working with LSEG Datastream data → read references/datastreamws.md
Working with NOAA climate/weather data → read references/noaa.md
Working with BEA, BLS, Census, or EIA data → read references/us-govt-data.md

Common Rules

Always use getSecret() for credentials and API keys — never hardcode secrets
Bloomberg, FRED, Haver, and Datastream require the Datafeed Toolbox.
fredrs additionally requires R2026a or later
datastreamws available since R2018b
NOAA is a standalone GitHub interface requiring MATLAB only.
US government data interfaces (usbea, usbls, uscensus, useia) are available on GitHub only and require MATLAB R2026b+ and Datafeed Toolbox R2026b+
Bloomberg and Haver local interfaces (bloomberg, blp, haver, haverdirect) are Windows-only; cloud/REST variants (bloombergHypermedia, bpipe, haverview, fredrs, datastreamws) are platform-independent

Files

ATTRIBUTION.mdAdded by AutoGineer1.0 KiB
LICENSELicence file added by AutoGineer1.4 KiB
SKILL.md6.6 KiB
manifest.yaml620 B
references/bloomberg.md11.5 KiB
references/datastreamws.md3.2 KiB
references/fred.md4.0 KiB
references/haver.md4.8 KiB
references/noaa.md5.8 KiB
references/us-govt-data.md8.2 KiB

references/noaa.md

NOAA Climate Data Online (CDO) — @noaa Interface

The @noaa class provides access to the NOAA Climate Data Online (CDO) Web Services API v2 from MATLAB. It is a GitHub-hosted interface, not part of the Datafeed Toolbox.

Installation

Clone the repository and add it to the MATLAB path:

% From a terminal:
%   git clone https://github.com/mathworks/noaa.git

% In MATLAB:
addpath("C:\work\noaa")   % adjust to your clone location

Requires MATLAB R2026b+.

Connecting

Create a connection using your CDO API token (register at ncdc.noaa.gov/cdo-web/token):

token = getSecret("NOAA_CDO_TOKEN");
c = noaa(token);

The constructor validates the token by issuing a GET request to the API base URL. If the token is invalid, it throws datafeed:noaa:connectFailure.

Discovery Methods

Six methods share the same calling pattern and all return tables (via struct2table):

[d, response] = methodName(c, id, Name1, Value1, ...)

id — optional record identifier; pass [] to omit
Name-value pairs map directly to CDO REST API query parameters
Default limit is 25 records per request; maximum is 1000 ("limit", "1000")
response is a matlab.net.http.ResponseMessage

datasets

d = datasets(c);                                  % all datasets (up to 25)
d = datasets(c, "GHCND");                         % specific dataset by ID

Returns table: uid, name, id (string); mindate, maxdate (datetime).

datacategories

d = datacategories(c);
d = datacategories(c, [], "limit", "42");         % all 42 categories

Returns table: name, id (string). No date columns.

datatypes

d = datatypes(c);
d = datatypes(c, [], "datasetid", "GHCND", "limit", "1000");
d = datatypes(c, [], "datacategoryid", "TEMP", "datasetid", "GSOM", "limit", "100");

Returns table: name, id (string); mindate, maxdate (datetime).

locationcategories

d = locationcategories(c);

Returns table: name, id (string). No date columns.

locations

d = locations(c);
d = locations(c, [], "datasetid", "GHCND", "locationcategoryid", "ST", "limit", "52");

Returns table: name, id (string); mindate, maxdate (datetime).

stations

d = stations(c);
d = stations(c, [], "datasetid", "GHCND", "locationid", "FIPS:12", "limit", "1000");
d = stations(c, "GHCND:USW00012839");            % specific station

Returns table: name, id, elevationUnit (string); mindate, maxdate (datetime).

Known issue: stations() can error on large result sets where station records have inconsistent fields. The struct2table call fails when some records have fields others lack. If this happens, filter by dataset, location, or limit to reduce the result set. For very large station queries, use direct HTTP requests with matlab.net.http as a fallback.

getdata — Retrieve Observations

[d, response] = getdata(c, datasetID, startdate, enddate, Name1, Value1, ...)

All four positional arguments are required:

Argument	Type	Notes
`datasetID`	string	e.g. `"GHCND"`, `"GSOM"`, `"GSOY"`
`startdate`	datetime or datenum
`enddate`	datetime or datenum

Common name-value parameters: stationid, datatypeid, locationid, limit, units.

[d, ~] = getdata(c, "GHCND", datetime(2023,1,1), datetime(2023,12,31), ...
    "stationid", "GHCND:USW00012839", ...
    "datatypeid", "TMAX", ...
    "units", "metric", ...
    "limit", "1000");

Returns a table with no column type conversion — all columns retain their JSON-deserialized types. This is different from discovery methods which convert specific columns to string and datetime. Typical columns: date (string), datatype (string), station (string), value (numeric).

To convert dates from getdata results:

dates = datetime(d.date, InputFormat="yyyy-MM-dd'T'HH:mm:ss");

Common Datasets and Data Types

Dataset	ID	Description
Daily Summaries	`"GHCND"`	Daily observations from ~100k stations worldwide
Monthly Summaries	`"GSOM"`	Monthly aggregates
Annual Summaries	`"GSOY"`	Annual aggregates

Data Type	ID	Unit (metric)
Average temperature	`"TAVG"`	tenths of °C
Maximum temperature	`"TMAX"`	tenths of °C
Minimum temperature	`"TMIN"`	tenths of °C
Precipitation	`"PRCP"`	tenths of mm
Snowfall	`"SNOW"`	mm

Year-by-Year Looping for Large Requests

The API returns at most 1000 records per request. For multi-year daily data, loop year by year and concatenate:

allData = table();
for yr = 2020:2023
    [d, ~] = getdata(c, "GHCND", datetime(yr,1,1), datetime(yr,12,31), ...
        "stationid", "GHCND:USW00012839", ...
        "datatypeid", "TMAX", ...
        "units", "metric", ...
        "limit", "1000");
    if istable(d) && ~isempty(d)
        allData = [allData; d]; %#ok<AGROW>
    end
end

Building a Timetable from Results

dates = datetime(allData.date, InputFormat="yyyy-MM-dd'T'HH:mm:ss");
values = allData.value / 10;    % tenths of °C to °C
tt = timetable(dates, values, VariableNames="TMAX_C");

Notes

The @noaa class is hosted on GitHub (github.com/mathworks/noaa), not shipped with the Datafeed Toolbox. It requires addpath after cloning.
Name-value parameter values should be passed as strings (e.g. "limit", "1000" not "limit", 1000).
Do not name variables datatypes or datasets — these shadow the @noaa method names.
Large API responses (1000+ items) may deserialize as cell arrays of structs instead of struct arrays, causing struct2table to fail. Reduce the result set with filters or limit.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Needs an API key

Security

No issues found by our static scan· No issues found

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by

MathWorks

Publisher

MathWorks (Official)

Source

matlab/matlab-agentic-toolkit · skills-catalog/computational-finance/matlab-access-datafeed · commit 2a7f215

Licence

BSD 2-Clause License (redistribution allowed)

LICENSE (added by AutoGineer from LICENSE.md)
ATTRIBUTION.md (written by AutoGineer)

Synced

2026-11-09

AutoGineer is not affiliated with or endorsed by MathWorks or MathWorks.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download matlab-access-datafeed.zip