A portal is not a single database
Public information about Spanish industry is distributed across several statistical operations and organisations. The Ministry of Industry and Tourism maintains an open-data area and topic-specific statistics pages; INE publishes official statistical operations, and Eurostat provides indicators that can be compared across European Union countries. Finding a figure on a ministry portal therefore does not mean that there is one database covering all industrial activity, or that every series can be combined directly. The first task is to identify which operation produces the data and what it measures.
The Ministry’s open-data page explains that the datasets made available can be used, reused and redistributed. This possibility of reuse is not the same as statistical completeness: the page alone does not establish that the Ministry publishes every company-level datum or industrial indicator, or that every resource offers a structured download, an API or regular updates. Before planning an analysis, open the specific resource and check its format, terms of use, responsible organisation and documentation. (Source: Ministry Open Data; electronic open-data portal).
The Ministry’s statistics and reports portal is an index of publications and series, not a guarantee of uniformity. According to the Ministry’s own description, its resources include downloadable materials in formats such as PDF, ODS and XLSX, as well as the BADASE series database. This combination can be useful for manual consultation and analysis, but it means distinguishing a published table from a historical series and from a methodology document. A download format is no substitute for a technical note. (Source: Ministry statistics and reports).
How to find and document a series
Start with the precise question: physical output, turnover, business confidence, employment, energy consumption and the number of companies are not interchangeable measures. Look for the name of the statistical operation and the exact table or series. Record its unit, geographical scope, reference period, frequency and publication or update date. If the page links only to a press release or summary table, check whether data can be downloaded and whether the methodology defines the variables precisely enough for your purpose.
For example, the Ministry publishes an Industrial Business Survey with sections for latest data, results and methodology. Its methodological dossier indicates that these surveys can provide rapid information about areas not covered by quantitative statistics. This means that a short-term business indicator should not automatically be read as a direct count of observed production: first check how its questions, population and construction are defined. Rapid publication can be valuable, but it does not make the results equivalent to quantitative indices. (Source: Industrial Business Survey; IBS methodology).
Keep the table URL, the date you accessed it, the available version or period, and any transformations you applied alongside every figure. For a download, also record the file name and, if the analysis must be reproducible, its checksum or a local copy. Avoid citing only the organisation’s homepage: readers should be able to retrace the path from the published claim to the table and its documentation. If data have been revised, note whether you use the updated series or an earlier release. A download date is not necessarily the statistical reference date.
Three kinds of indicator, three different questions
INE’s Industrial Production Index (IPI) measures changes in industrial production, according to the description of the statistical operation. Its published methodology reports monthly data collection and a sample of industrial establishments; the scope and calculation should be checked in the current documentation for the operation. The IPI is useful for analysing production changes, but it is not, by itself, a measure of tonnes, units manufactured or sales value. Nor does it automatically describe every establishment, product or branch at the same level of detail. (Source: IPI operation and latest data; IPI methodology).
Industrial turnover indices answer a different question: they track changes in the turnover reported by surveyed establishments and are presented as indices, with a base year defined in the methodology. An increase in the index should not simply be described as the same percentage increase in units produced or in nominal revenue for every company. The operation, base, adjustments and comparison period matter. This is an aggregate indicator of commercial activity, not an individual company’s income statement. (Source: INE metadata on industrial turnover indices).
Industrial business confidence, by contrast, provides signals about business perceptions and expectations, in line with the purpose set out in its methodology. It can complement quantitative indicators and help put changes in context, but it does not replace them. Vehicle or cement production is another example: these are sector-specific statistics and, by definition, should not be presented as a picture of the entire industry. The Ministry’s vehicle manufacturing page links to specific series and reports, while the cement statistics state that their purpose is to track the cement sector. (Source: vehicle statistics; cement statistics).
Cross-checking sources without inventing equivalences
Eurostat provides a series for total industrial production excluding construction, as well as metadata for European industrial production statistics. Its metadata states that Spanish methodological documentation is published by INE and that data are transmitted to Eurostat in SDMX format. The European series can help compare countries or consult a harmonised presentation, but before matching it to a Ministry table, check the definition, adjustment, unit and period. The shared label “industrial production” is not enough to prove that two tables are identical. (Source: Eurostat, industrial production; Eurostat metadata).
Company indicators also require particular care. The Ministry’s SME platform describes its tables as based on information from INE’s Central Business Register (DIRCE), with variables such as size, legal form, activity, revenue, and geographical and time scope. The Ministry also describes “Cifras PYME” as a monthly publication based on data about companies registered with Social Security. These sources are useful for different questions, but do not assume that DIRCE and Social Security registration records count the same units or use the same activity criteria. (Source: data on Spanish companies; SME statistics and publications).
When comparing series, align their frequency and reference period before calculating differences. Check whether a figure is monthly or annual, an index, a physical volume or a company classification; also check whether the series is adjusted and whether its base year or methodology has changed. If two sources diverge, do not automatically choose the newest or simplest figure: explain the conceptual difference or avoid the comparison. A valid comparison begins by showing that the objects being compared are sufficiently alike.
Limits worth making visible
Portals and results pages can change, and a page linking to data does not always expose the full structure, frequency or revision history through its search interface. The evidence consulted can identify resources and some methodological features, but it does not justify claiming that every dataset is open in reusable formats, that a particular download is updated at a stated frequency, or that every operation has an API. Check those details on the specific page and file before using them in a project.
The level of detail also matters. A national total can conceal differences between branches, territories and company sizes; a sector statistic does not describe other industries; and a sample is not a census. Conversely, a detailed table may not be precise enough for every cross-tabulation. If the documentation does not support a breakdown, do not reconstruct it through inference or present it as an observed datum. The absence of a published detail does not prove that it does not exist in other records; it marks only the limits of the source consulted.
Interpretation must respect the uncertainty and purpose of the operation. An index summarises changes according to a methodology, not the condition of every company; an opinion survey records responses under a particular design; an annual figure may arrive later than a monthly short-term signal. There is no universally “best” source: the right source depends on the question, and any article citing it should say what it measures and what it leaves out.
A short checklist before citing
Before publishing an industrial figure, state five checks in plain language: who produces it; the exact operation and table; the unit and population it covers; the period it refers to; and where its methodology is described. Add a sixth if you are comparing figures: are the definitions, adjustments, frequency and coverage compatible? If you cannot answer, rephrase the claim or explicitly limit its scope.
In a technical or business note, cite the direct link to the series and its methodology, state the access date and reproduce the unit as it appears in the source. Distinguish the observed datum from your interpretation. For example, do not turn an index into physical volume or a sector publication into a conclusion about the entire industrial fabric. If you use several sources, explain why they complement one another and which differences prevent treating them as a single series.
Spain’s public-data ecosystem does provide useful starting points: the Ministry for certain publications and sector statistics, INE for national statistical operations, and Eurostat for European data and metadata. Their existence alone does not show that every indicator is available at the same level of detail, on the same schedule or under the same licence, or that all series can be compared without adjustment. Verification is not just finding a figure; it is preserving the conditions that give it meaning.