• Home
  • Careers > Diverse Careers following an education at ENSAI

Diverse Careers

Today, information, be it quantitative or qualitative, is the key element in all decision making, and because of this, it has never generated so much talk. The science of statistics allows us to give objective answers to the strategic questions that economic and political players ask themselves.

Lets look at a few examples of the diversity and importance of the work that can be carried out by statisticians.

> Complementary information:
Newsletter # 46, #48, #50 and  #52 of the Ensai on the professional insertion of students
Zoom et vidéo sur les Métiers de la Statistique réalisé  par la SFdS en partenariat avec l’Onisep
Brochure of the American Statistical Association
Statistics jobs presented by the French Society of Statistics (SFdS) (more information)
Report on the 2013 survey of the CGE on the profesional insertion of graduates.
Website deviensingenieur.fr by the CDEFI
Le commerce en ligne français s'arrache les « data miners » Lire l'article des Echos
Internet recrute les boss des maths> Lire l'article de Ouest France
Why Statistics?  > Lire l'article
Big Data, gros enjeux  > Lire le billet de Gilles Babinet paru dans Les Echos le 20/07/12
Devenir fonctionnaire : le parcours de Justine, statisticienne à l'Inse> Lire l'article de l'Etudiant

Hugo Varet, statisticien à l'institut pasteur - Ensai Hugo Varet, statisticien à l’Institut Pasteur Fabien Domergue, Analyste revenue management à Disneyland Paris - Ensai Fabien Domergue, analyste revenue management à Disneyland Paris Audrey Pichavant, ingénieur chercheur EDF Recherche et développement - Ensai Audrey Pichavant, ingénieur chercheur en statistique, EDF Recherche et Développement 


Risk Management Statistician in the banking sector

Over the last 20 years or so, risk management has gained an essential position in banks. The Bâle committee, which gathers the central banks of the G10, has developed a series of banking norms in order to understand and deal with risks in banking and in this way, to constitute certain demands concerning capital to avoid the risk of going bankrupt and having this spread to the economy as a whole. Financial crises have shown the catastrophic impact that banks filing for bankruptcy can have on the whole economy.
This set of rules requires very precise risk management. And this is where the skills of an ENSAI statistician,  who is specialized in banking regulations and the quantitative techniques associated with them, become very useful. Indeed, banking risks require in-depth knowledge of innovative methods such as the theory of extreme values or the copula theory in order to efficiently measure the risks taken by banks. Risks are linked to when banks lend money to private individuals or companies (credit risk or the borrower fails to pay back), to market activity (market risk or financial asset fluctuations risk) or  simply to transactions: errors, fraud, etc (operational risk). Statisticians use all their skills in computing data and  statistics to establish, for example, an equity ratio that is in line with regulations.

Financial Asset Management Statistician

The management of financial assets is currently experiencing an important gain in popularity that is unlikely to stop. Indeed, the problem of retirement funds in our increasingly globalized economies calls for statisticians who are able to avoid savers from losing money, and more importantly, to help them earn money.
In order to fulfill this objective, financial institutions that deal with savings need statisticians who can analyze financial market data accutely to achieve the highest financial gain for the saver. A very competitive environment has now set in, to try and have the best investment options and create performance. ENSAI statisticians are able to use the most recent portfolio management techniques and the most innovative quantitative strategies (Kalman filter, ACP, etc.). ENSAI statisticians have solid knowledge of financial markets and of the economy, which is essential for investing intelligently in the markets.


Customer Relationship Management

Today, banks face increasing competition and their "clientele" strategy must be adapted to the new conditions that apply to the banking market. The cost  of "recruiting" new customers is such that it is sometimes more profitable to gain the loyalty of existing customers than to try and "conquer" new ones. CRM (Customer Relationship Management) policies that are being developed massively in the banking sector, like in many other sectors, have the goal of leading the marketing operations needed to gain the loyalty of existing customers. However, these policies require mobilizing an important amount of the information that a bank has on its own customers. In order to correctly anticipate a customer's decision to leave the bank, it must first gather and analyze all the information on that customer. And here, statistical skills are what are needed. Studying the attrition phenomenon (risk of not gaining the loyalty of customers) is not only useful to describe and  understand the behavior of customers who may leave the bank, but also to give a probability that this event  might occur and if possible grasp the result. The attrition score will have to answer this triple objective to allow the bank to react in time to "counter" this departure. By taking into account the value of the attrition score, several marketing operators can then be led toward these customers.

Geomarketing Statistician

In France, the first geomarketing studies date back to the late 1980s. Since, the utility of these studies has never been greater and their use is increasingly popular. By superposition of a grid for reading territories and a grid for reading consumers, geomarketing allows us to know the local potential for consumption and the differences in consumer behavior in order to segment demand and target commercial action. Today, most companies in the banking, insurance, retail, transportation, or automobile industries use geomarketing. The evident simplicity of the concepts is only apparent: it is a highly technical area where data is computed, using the keywords: supplying, managing, structuring, and representation of data. Relationship marketing and geomarketing are now modern concepts of businesses on their markets. Modern consumers are more demanding, more alert and require individualized attention. Adapting the offer to local needs, personalizing the services that come with selling, and the direct relationship with the customer are now the favored means of developing market shares. Furthermore, the demand for geomarketing will undoubtably be stimulated by the new developments in computer technology, namely the development, the availability and the improved quality of data bases offering increasingly diversified and precise views on behavior, thus making this new approach more attractive.

Other examples:

- Analysis of consumer demand for products and services
- Establishing a population – target of a marketing campaign
- Design and launch of a service to help marketing via online sales
- Representation on a model of target groups using data mining
- Positioning brands
- Creation of appetence scores
- Market studies, structural analysis of competition


Epidemiologist Statistician

Epidemiology is the study of links that exist between pathologies or any other biological phenomena and diverse socio-environmental factors (lifestyle, social or ambiant circle, individual characteristics…) that can have an influence on their frequency, their distribution, their evolution… This science is very much based on statistical methodology. Many organizations, and more particularly INSERM, have created research units in epidemiology in which many statisticians work.
The sometimes tragical events linked to mad cow's desease, to dioxin in chickens, to listeriosis… have incited health organizations to massively develop health risk evaluation studies in both human and animal epidemiology. In other sectors, such as the environment, the risks linked to pollution are also very important and can have consequences on health. For example, the regular rise in the number of incinerators linked to the rise in the quantity of household waste over the past 20 years, has led organizations to request an evaluation of the risk of  congenital deformities of masculine and feminine origin linked to emissions such as metals, dust and dioxins. This particularly complex work, because of the "time" but also taking into account the differences of situation between exposed and non-exposed zones, can lead to confusion: extra road traffic adding to the "dioxin," industrial areas that make the incinerator effect rather "fuzzy"… statisticians need to be aware of all these factors to conclude their study rigorously and bring non-contestable answers.


When a new drug is being developed by a pharmaceutical laboratory, biostatisticians have the task of collaborating with clinicians throughout clinical testing from the development of the protocol through to data analysis. The authorization to market a new pharmaceutical product requires submitting, to the regulatory authorities (in France it is the Agence Française de Sécurité Sanitaire des Produits de Santé, or AFSSAPS), a technical application that contains 3 development phases. Generally, statisticians come into play from the first phase (administering the drugs on test subjects) and the analysis of the first tests on humans. The data (socio-demographic, biological, clinical, semeiological...) gathered using observation books used by the practionners are studied in the biostatistics department after management by the data-management department. The results are then put into a report and given to the clinical research team of the laboratory. The result of this work is very concrete and valorizing because it results in the launch of a new drug on the market.
> Anne-Cécile Bourien, biostatistician

Other examples:

- Analysis of the impact on the state of daily drug consumption (alcohol, tabacco, cannabis…)
- Validation of a quality of life scale for chronically ill people
- Evaluation of the risk of congenital deformities for pregnant women living near an incinerator for household waste
- Creation of a method to detect genes that slow the progress of an illness
- Biostatistical researcher
- Researcher in genomics


Statistician in Sensory Analysis

Very often, consumers are unable to explain the reasons why they prefer one product to another. It is very common for consumers to like or dislike a product, whether moderately or passionately, without being able to make the choice legitimate or justify their appreciation of the aspect, the feel, the odor or the taste of a food product, for example. That is why the food industry, cosmetics companies, automobile companies… are increasingly using the skills of sensory analists. R&D departments of companies developing sensory analysis in order to get descriptions of the products to be produced and marketing services have preferential data on these products. Having techniques that allow the bringing together of the two approaches by joining together the two information groups represents significant progress in adapting the products to the market. The mapping of the preferences is the statistical tool that allows for this union.
Broadly, the method consists firstly in creating sensory profiles for the products using expert panels specifically trained to presicely and completely describe the products and secondly, in interviewing some consumers who represent the target population on their level of appreciation of these same products. The fusion of these two sets of data and their analysis allows for the explanation, in objective and quantifiable terms, for what justifies the consumer's attraction or rejection of a product. With this mapping of preferences, R&D departments can make their products evolve towards one group of consumers or another.

Other example:

- Following the quality of an agronomic production


Industrial Statistician

Safety is a must in the industrial world. The prevention of unwanted events in a production process is an important field in statistics: reliability. Reliability engineers, trained in statistics, must understand (how does it work and how does it break?), model and compute (the probability of non-desired events taking place), identify and class (the scenarios and actions that are the most effcient in terms of profitability). The keystone of this entire system is made up of basic numerical data (deficiency rates, repair rates…) gathered thanks to experiments and used in probabilistic models. It must, however, be recognized that the data gathered from safety studies are sparce and incomplete. Thus, it becomes essential to use methods that are adapted to safety problems like dealing with missing data, using Bayesian methods, collecting data in the life cycle… Evaluating the probable future linked to acquired knowledge from the past, such is the quest of the reliability statistician.

Statistician in Computer Image Processing

A developement pole where the skills of statisticians are particularly important is research in computer image processing, whether in the academic field or in the industrial world. The earliest work in this field (automatic shape recognition, text recognition, movement analysis, etc.) called for simple solutions and initiatives where the initial skills required were mainly computer related. These skills soon became insufficient when faced with the described models that still require computerized testing. Thus, over the past few years, the processing of image analysis developed around Markov models (Markov fields in analysis of the apparent movement or in image restoring, hidden Markov chains for recognizing handwritten text), around factor analysis (face and symbol recognition, etc.), or around image compression (for digital television, for example). This last method is based around the statistical analysis of results that come from the theory of information so as to only release useful and predominant results (entropic coding, minimum description length). In these fields, strong statistical skills are now required to fully understand these tools. Also, when these skills are combined with an in-depth knowledge of computer science, this profile becomes particularly attractive to research laboratories that deal with image processing.
This field of work is also accessible through the 3rd-year Statistical and Decision-Making Systems specialization.

Other examples:

- Study of the operating safety of a nuclear power plant
- Evaluation of the consequences on the environment, especially air and water quality and excessive use of pesticides
- Detection and establishment of a target in the case of numerous false alarms
- Forecasting ozone peaks in urban areas
- Estimating the life span of a car air conditioner

Computer Science

Computer statistician

Engineers processes large amounts of data by computer. They can use, adapt or even create their own statistical analysis tools. The professions that best correspond are datamining and the management of datawarehouses. These professions allow them to use skills on both statistics and computer science in their company, in any application area (personal date, bank data, bio, etc.).

  Other examples:

- Creation of a computer tool to validate and follow scores;
- Design and implementation of an automatic video flow classifier.

Civil Servant Statisticians for INSEE

The classes taught at ENSAI prepare civil servant statisticians for INSEE to exerce, in the French Public Statistics System (SSP), a host of technical and scientific professions centered around statistical engineering and information systems. Thus, they take part in the production and utilization of the national population census, in carrying out surveys concerning specific populations (households, businesses, unemployed people, high school students…) targeted according to the interest they represent in analysing the socio-economic situation in France and its territories. They carry out the statistical processing, the analysis and the publication of results. Beyond surveys, the use of administrative sources allows for bettre knowledge of the economy. It is the case for example when dealing with tax revenues or when looking at the flow of labor in companies. They can also be called to help in the creation of the national accounts, to create conjunctural statistics or apply the necessary computer knowledge for the activities of the SSP.

Other posts can also be offered to them at INSEE or in the statistics services of different Ministries such as Research, Management, Communication, Marketing, Commercialization...

Devenir fonctionnaire : le parcours de Justine, statisticienne à l'Insee  > Lire l'article de l'Etudiant