Working With Spanish Labor Market Terminology Right Now
Most people coming from outside Spain treat labor market terminology as a simple translation problem. It isn't. The vocabulary has shifted significantly over the last three years, and the official categories from the Instituto Nacional de Estadística don't match what employers are actually writing in their postings. This mismatch creates real friction when you're trying to build something that maps candidates to roles or when you're analyzing hiring trends programmatically. I ran into this exact problem last year while working on a project that needed to tag job postings by sector and seniority level across multiple Spanish provinces. The dataset came from a public scraping effort, and the labels were inconsistent by design. Some postings used "desarrollador," others used "programador," and a growing number used English titles like "software engineer" even in fully Spanish-language ads. The official INE occupational taxonomy (CNO-11) uses "Informáticos y desarrolladores de software" as a category, but nobody on the actual job boards uses that phrase. If you map directly against the official taxonomy, your precision drops to roughly 42 percent. The workaround I settled on was building a lightweight normalization layer that maps common posting titles to standardized codes. I used a combination of exact string matching against a curated synonym list, fuzzy matching with a threshold of 0.82 on Levenshtein distance, and then a small rule-based fallback that checks location and required experience fields to disambiguate cases where the title alone is too vague. The whole pipeline runs in about 8 seconds per 10,000 records on a standard container. Not fast enough for real-time use, but fine for batch processing overnight.
One detail that nobody warns you about: the term "autónomo" appears in job contexts in three completely different ways. It can mean a self-employed contractor position, it can refer to the legal status of the company posting the ad, and it sometimes appears in job descriptions as a benefit clause ("posibilidad de trabajar como autónomo"). If your parser treats every instance as the same thing, your classification of contract type is going to be garbage. I solved this by checking the semantic context window around the word. If "contrato" or "baja por" appears nearby, it's a status flag. If "como" or "modalidad" is in the same sentence, it's a work arrangement. Three minutes of manual annotation on 500 examples was enough to train a tiny classifier that handles 94 percent of cases correctly. Another thing worth noting is the regional variation. The Basque Country, Catalonia, and Galicia all have parallel terminology in their local languages that doesn't always have a one-to-one mapping back to Spanish. Tools that only process Spanish will miss or misclassify a significant chunk of postings from those regions. I added a simple language detection step using langdetect and routed Basque and Catalan postings through a separate synonym dictionary before the main normalization pipeline. This improved overall coverage by about 11 percent, which sounds small but matters a lot when you're doing regional hiring analysis. The main limitation of this approach is that it doesn't handle novelty well. When new job titles emerge — and they do, especially in tech — your synonym list becomes stale within months. I've found that maintaining an automatic suggestion queue where flagged low-confidence matches get reviewed weekly keeps the system usable without requiring constant manual updates. The queue itself takes maybe two hours per week for someone who knows the domain.
There are commercial platforms that claim to solve this problem out of the box, but they charge per API call and their customization options are thin. The approach above is cheap, transparent, and you own the mapping logic. If you need something more sophisticated for production at scale, there are specialized labor market data providers that maintain updated taxonomies, but they cost several thousand euros annually and still require some local tuning for the Spanish market specifically.
Get the Full Details
