Today, billions of items of data are being made accessible via the web, served to a variety of conventional devices (PCs, tablets, smartphones) and connected objects which are either private or made available to a wider audience as Open Data. For those seeking to go beyond the raw data and make sense of this new “black gold”, semantic analysis is an essential building block for managing data, enabling a harmonised (and even standardised) approach.
Intelligent data use calls for understanding, interpretation and classification
Modern data is diffuse in nature because of the multifarious information sources from which it derives. Quantitative data may be standardised, but in a way which draws upon different frames of reference; these need to be reconciled, and even harmonised. Furthermore, up until now, qualitative data in text form have rarely (if ever) been structured, and remain unexploited. These rich textual data sources are almost impossible to reconcile without a suitable tool. The latest generation of tools, exploiting semantic analysis methods, can offer new analysis insights.
Semantic analysis can take us far beyond traditional search engines
Semantic analysis first appeared in 1883 in the work of Michel Bréal, but really came into its own in the 1980s with the development of cognitive sciences. The strength of semantic analysis is its ability to transcend the limits of a traditional search engine. For example, in a search engine, if you enter “energy company country” to find companies that are active in energy in different countries, the search will return a series of articles containing variously stronger or weaker combinations of the three words. With semantic analysis, your search takes on a whole new meaning. The words “energy company” and “country” will be converted into concepts. And the displayed hits will be articles which include companies from the energy sector, with information on their location. To give another example, semantic analysis can create on-the-fly tags to be attached to a document in order to improve information retrieval in an EDM (Electronic Document Management) system.
By reading and understanding the meaning of text drawn from multiple sources (tweets on Twitter, comments on social networks, email content, etc.), it may also be able, for example, to detect moods, opinions or levels of satisfaction. Goals such as detecting trends and consumer expectations, highlighting fashionable topics, monitoring marketplaces and competitors are all convincingly and easily within reach without the need for lengthy, expensive market research. Combined with voice recognition – already widely available on our smartphones – it can transcribe and interpret voice recordings. With its ability to operate in multiple languages, it will push the limits of our linguistic barriers.
New perspectives for Human Resources
Many IT companies are positioning themselves in this market, each with its own interesting DNA: some are generalists, while others have opted for the vertical specialisation route. And some players have decided to focus on HR business processes, offering HR document analysis tools (e.g. job descriptions, CVs, competency dictionaries, job offers, career paths, training courses, etc.). With this technology, a wide range of possibilities opens up to HR. Imagine what we could be doing with the wealth of information inside all those assessment interview form text fields! While this data has until now been entered mainly for information purposes rather than for exploitation, it could in future be used, for example, for improved analysis of the social climate using so-called “weak signals”, for better identification of unmapped skills in companies’ formal frameworks, and for fine-tuning individual and collective performance, which are all too often measured using necessarily reductive quantitative indicators. Just think what could be done with a tool which was able to collect, interpret, structure and analyse data hitherto stored passively inside HR information systems, project management systems, collaborative tools, corporate instant messaging systems and emails… not to mention sources outside the company such as professional networks, job boards, online training catalogues, and HR data in Open Data format on the web.
However, with each of these information systems possessing its own languages and frameworks, cross-referencing their content quickly becomes an impossible task without preparation and standardisation work. In this respect too, semantic analysis tools offer a very efficient solution for harmonising (and thus processing) all this data. In this way, these tools could change the way in which HR frames of reference (skills, remuneration, positions, training, etc.) are built. Existing frames of reference are the result of long, tedious preparation work, and generally tend to reflect past realities instead of analysing future developments. With semantic analysis tools, they can become more responsive, forward-looking and practically useful. They offer a genuine solution for transforming companies.
Semantic analysis is a method which has been known about for a long time, but has frequently been under-utilised. The latest developments in IT, combined with the power of associated technology platforms (hardware & middleware), have made such resources available to everyone for a range of increasingly relevant uses, including HR. Lastly, semantic analysis offers the prospect of shifting HR information systems away from an approach which has usually been merely quantitative and standardised, instead providing qualitative solutions in which norms and standards will no longer be reductive shackles, but rather agents for harmonising accounting practices and adding muscle to analyses. Built dynamically to work closely alongside employees and managers, and enriched with data from outside the company thanks to Big Data technology, they will be living, rich resources which can be easily and seamlessly updated.