How to map commodity producers and traders from job data

To map who produces, trades and buys a commodity, filter job posts on the commodity's name, keep the companies behind them, and split those companies by the work their posts describe and the countries they hire in. Companies that liquefy, ship or trade a commodity write it into their job ads: operators for the plant, traders for the desk, terminal staff for the import side. In TheirStack the commodity is a buying intent topic, so one filter finds every company whose posts mention it, and each match comes with the posts that put it there.

As of 2026-11-03, 6,719 direct employers have mentioned liquefied natural gas (LNG) in their job posts since we started tracking it, and 893 of them have open LNG jobs posted in the last 30 days (2,693 jobs). The rest of this playbook builds that LNG map step by step. The same steps work for any of the 83 commodity topics we track, from crude oil and maize to polypropylene.

Why do job posts show who produces, trades and buys a commodity?Copy Anchor Link

A company names a commodity in a job ad when the job handles it: a plant operator on an LNG train, a trader who books cargoes, a technician at a regasification terminal. Every role it posts adds to a record of what it does with the commodity. Trade directories and member lists tell you who exists; job posts tell you who is staffing up around the commodity right now.

We detect commodity names in every job we collect and roll them up to the company. The buying intent data docs list Commodity as one of the 14 keyword types, next to regulations, equipment and strategic initiatives.

Who needs a commodity market map?Copy Anchor Link

Analysts who need a market map as the output: a procurement or supply-chain team sizing the supply side before a sourcing decision, a market-research or consulting team building a landscape slide, an investment analyst looking for who is expanding capacity. Here the list is the deliverable, and nobody on it gets an email.

How do you build the LNG search in the app?Copy Anchor Link

Start in Job search, because a market map needs the posts as evidence and each row there is a post you can open.

1. Click Add filter and pick Buying intent topic. Type the commodity (here, "Liquified Natural Gas (LNG)").
2. Set Date posted to the last 30 days, or to "all time" for a historical map. Our job history starts in 2021.
3. Set Job Closed to show only open jobs.
4. Set Company type to Direct employer, so recruiting agencies don't count as producers.

Copy Text

Buying intent topic: Liquified Natural Gas (LNG)
Date posted: last 30 days
Job Closed: open jobs only
Company type: Direct employer

Open this searchOpen in ClaudeOpen in ChatGPTOpen in CursorConnect TheirStack MCP

That gives you the whole market. To read it as a supply chain, add one of these filters at a time to the whole-market search and save each as its own search. Each one is a row in the table below:

• Producers: Job description matches (?i)\bliquefaction\b|\bLNG trains?\b|\bexport terminal\b.
• Plant roles: Job title matches \b(operator|technician|process engineer|plant|liquefaction|production)\b. It narrows the whole market, not the producer slice, so it can return more companies than the producer row.
• Traders: Job description matches (?i)\b(LNG|gas) trading\b|\bcargo(es)?\b.{0,40}\b(trad|portfolio)|\bportfolio optimi[sz]ation\b.
• Trading desk roles: Job title matches \b(trader|trading|origination|originator)\b.
• Buyers: Job description matches (?i)\bregasification\b|\bregas\b|\bFSRU\b|\bimport terminal\b.
• Utilities: Company industry is Utilities.
• Procurement roles: Job title matches \b(procurement|buyer|purchasing|sourcing)\b. Read these posts before you count them as buyers (see the pitfalls below).

For the geography, add Job country one region at a time. Job country is where the role is based; use Company HQ country instead if you want the company's home market.

When you want companies instead of posts, run the same filters in Company search: its job posting filters return each company once, with the number of jobs matching your filters, and the Buying intent topic filter adds a column with our confidence and the first and last date we saw the topic at that company.

To size a search through the API before you spend anything on it, call Company search with include_total_results: true and limit: 1. A count costs one record (3 API credits for a company), and a search with no matches costs nothing, as the counting results docs explain:

curl -X POST 'https://api.theirstack.com/v1/companies/search' \
  -H 'Authorization: Bearer <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "company_keyword_slug_or": ["liquified-natural-gas-lng"],
    "company_type": "direct_employer",
    "include_total_results": true,
    "limit": 1
  }'

As written, this returns every company that has ever mentioned the topic (6,719 on 2026-11-03). Add job_filters with a date filter to keep only companies posting about it recently. The MCP tools page lists the tools an agent uses to run the same searches:

Using the TheirStack MCP, find the direct employers with open job posts from the last 30 days that mention the buying intent topic Liquified Natural Gas (LNG). Split them into producers, traders and buyers with the job description and job title patterns from this playbook, count the jobs and companies in each slice, and show five job posts per slice so I can check the matches.

Open in ClaudeOpen in ChatGPTOpen in CursorConnect TheirStack MCP

How many companies hire around LNG right now?Copy Anchor Link

LNG jobs posted in the last 30 days, open, from direct employers, as of 2026-11-03:

By where the jobs are based:

Slices overlap: an operator post that mentions an export terminal counts in both the producer and plant-role rows, and a company hiring in two regions counts in both. Read each row as its own list, and don't add them up.

On the producer side we found a plant operator for an LNG facility in the US, a process engineer for LNG in Indonesia, and a senior operator for LNG trains and helium plants in Qatar. The trader slice brought up a senior cross-commodities and power trader at a gas marketing and trading company in London and a market risk manager on a German utility's energy trading desk. On the buyer side, a German import-terminal operator was hiring a stakeholder manager for its FSRU sites, and a shipping line in Hong Kong was hiring an officer for its FSRU.

How do you turn the list into a market map?Copy Anchor Link

For a map, working the list means reading it, cleaning it and keeping it current.

Read the posts before you trust a slice. Open five matches per row: if they describe the role you meant, keep the filter; if they don't, tighten the pattern. Then reveal the companies and export the list (each reveal costs 1 company credit and stays visible for 90 days). The company export includes the firmographics, the number of jobs matching your filters and, for each topic you filtered by, its confidence, job count and first and last date seen. If you'd rather load the raw records into your own tools, datasets deliver jobs and companies as CSV or Parquet files, each with a data dictionary.

Deduplicate the file by domain. We don't match companies on their domain alone: we index them as entities and merge the records of one employer from several signals, so the same business can still appear twice under slightly different names that share one website.

Label each company with its role in the chain, using the slice it came from plus your own reading of its posts. A company in both the producer and trader slices may be an integrated player; keep it in both and flag it.

Refresh on a schedule. Save each slice and turn on email alerts or a webhook so new entrants reach you as they start hiring, and rerun the counts each quarter to see which part of the chain is growing.

Where can a commodity map go wrong?Copy Anchor Link

• Service providers mention the commodity too. Engineering contractors, inspection firms, consultancies and software teams name LNG because their client works with it. In the procurement-roles slice (43 jobs at 29 companies as of 2026-11-03), the matches we read were contractors buying equipment for LNG projects: they belong on a map of the project supply chain, not on a list of LNG buyers. A consultancy implementing commodity trading systems also showed up in the trader slice.
• Check the company record. In the 20 posts we read for this playbook, an expediting role for LNG project equipment sat under a sister company filed under retail apparel. Before a company goes on a published map, check that its domain and industry match the employer the post describes.
• Agencies can slip through. Direct employer removes companies we classify as recruiting agencies, but a small agency posting a single "LNG Trader" role can still look like an employer. Read its company profile before you count it as a trader.
• Lists of industries count as mentions. Some posts name LNG only in a list of sectors a candidate may have worked in, next to refining and pharmaceuticals. Those companies have nothing to do with LNG, so a slice built from a free-text pattern needs the same five-post check as the rest.
• Hiring is a proxy. A producer that is fully staffed posts little, so job counts show who is active and growing; they don't measure production or market share.
• Thin commodities give thin maps. A niche commodity may only have a few dozen companies hiring in a month. Widen the date range, and say how many posts sit behind each number when you publish it.

Open the LNG search and swap in your commodityCopy Anchor Link

Open the LNG company search in TheirStack: direct employers with open LNG jobs posted in the last 30 days. Swap the topic for your commodity, then add the role and country filters above. The same filters as a job search show the posts behind every company.