Compu-Graphics to Visual Analytics: 80s and after

After the Indo-French Compu-Graphics and Planning Project was over in 1985, I moved back to France as a Technical Counsellor to the Cabinet of the Minister for Cooperation, to develop, with Luc de Golbéry, a Visual Decision Support Information System in “Cellule de Synthèse et d’Information Rapide – CSIR”[1], a Unit for Synthesis and Rapid Information, to help the Minister’s counsellors analyse information about Franco-African cooperation. Only to discover how much graphic processing and maps – neutral, possibly exhaustive and transparent – disturb the ministerial antechambers accustomed to discretion. We were asked to not show and circulate our documents. Visual output was considered dangerous, too effective?

French Public Aid: How Much and Where? Plotter output

After the change in government the information cell was dismantled. I was offered a post in the Ministry of Foreign Affairs, but doing something else than analysing and visualising data. I realised that what interested me in life was analysing data with a purpose of serving the most deprived. I decided to start my own company, Decision Graphics, in Paris. I did all the subsequent work from that structure till 2005, when I create a company in India, in Hyderabad, GMIS Consultants.

UNESCO asked me to participate in their training package “Government Information Systems and Socio-Economic Data”, using the materials developed in India and in Sri Lanka to prepare the module 6 (Maps and Graphics: Tools for Statistical Processing) and 7 (Geocoding and Information Systems) for the Intergovernmental programme for co-operation in the field of scientific and technological information, UNISIST- World Science Information System.

Thierry Gaudin, head of the “Centre de Prospective et d’Evaluation”, Ministère de la Recherche et de la Technologie, France, asked me to build a world data base from 1960 to 1985 with projections to 2100 and prepare graphic analysis and outputs for the publication “2100, récit du prochain siècle, Payot, Paris 1990”[2]. This experience, working with challenging young graphic designers, changed my way of preparing visual output, from academic black and white to using colours in a very unconventional way for academics.

Initial “academic output
Final output in the book

Between 1990 and 1997, I worked in France with different local bodies, developing decentralized micro-GMIS for local level democratic management, planning and decision making. The scale varied from a region (Normandie in France: 3250 villages and towns, 240 “cantons” similar to mandals/teshils, 5 districts) to a group of 100 “communes” or panchayats in 6 “cantons”. The work was an iterative one, involving preparation of visual analysis and communication documents presenting a spatial diagnosis to help build future scenarios, that were presented in group discussion with the local stakeholders. This exercise implied simplifying and communicating complex information, without over-simplifying and caricaturising, and with enough clarity to spark discussions. This experience was partly successful, at local level, with the group of 100 “communes” who used the visual output in different meetings. I recollect one particular instance when there was a heated discussion for close to one hour between two communes before they finally came to discuss scientific arguments and action plan. The power of maps is to make visible and unmistakable issues that are otherwise hidden under emotional layers of untold resentment. But at a higher level the reception was colder. The maps and other visual output were too telling, hence judged dangerous, and I was asked to consider them strictly confidential[3].

The controversial maps showing contrasted evolution and reversal trends between two communes

In 1996, I met Mr Smarajit Ray, IAS, Principal Secretary in charge of Social Welfare, and I proposed to apply the methodology and experience developed with local bodies in France to the Ministry of Tribal Welfare and Girijan Cooperative Corporation (GCC) of Andhra Pradesh. We applied it to monitor Public Distribution System, Minor Forest Products procurement and Agricultural Loans, through an IFAD project. It brought to light the delays and the Manager could focus on the areas of concern more quickly. The tool developed was combining GIS (geographical information system) and MIS (management information system so I coined the term GMIS combining both into a single system. The Managing Director of GCC, Suresh Chanda, IAS, was keen to show the tool to Chandra Babu Naidu, then chief minister. This proved to be a disaster, as the CM, instead of focusing on the tool and how it would be used, focused on the particular delays and mistakes visible on the maps and was very critical of the IAS officer in front of every one. This discouraged him to pursue the experiment.

My work was then noticed by the IAS secretary of the A.P. Social Welfare Residential Educational Institutions Society. He asked me to develop a visual tool to monitor school infrastructure, school performances, building of schools, location of new schools. For the latter, the A.P. government had sanctioned the creation of 40 new boarding schools. The question was a) to distribute between boys and girls and b) where to locate them. It was assumed that there should be a) an equal number of schools for boys and for girls and b) an equal number per district to ensure equity. But after analysing a) the literacy level, and particularly the social and gender gaps in literacy, and b) the location of existing schools, it was decided to correct the noticed imbalances. It was decided a) to create mostly schools for girls and b) to create more schools in Telangana where the existing schools were much less. To get the approval from the CM the administrator prepared the file with maps and gave them to the CM just before he left for a helicopter trip that would allow him to spend one hour free of calls and disturbance. The CM came out of the helicopter approving of the project. And every time I pass in front of the SC Residential School of Dindi, a very poor area, I feel proud to have contributed to its location, that would otherwise have been missed. [4]

Female Literacy Rates

Back in France I worked on a project for restoring the maritime character of Mont-Saint-Michel, registered since 1979 on the UNESCO world heritage list, located between Bretagne and Normandie, for the “Agence de l’Eau”, Water Management Board, and “Mission du Mont-Saint-Michel”, Ministère de l’Equipement (R&B). I helped prepare atlases for the five water basins around Mount Saint-Michel. 1997-2000

Teaching in Master TRIAD 2002 and in India different projects.

In 1999, my husband, Luc de Golbéry, was contacted by a senior IT consultant, Jean-François Rossato, who had visited us in Hyderabad during the Indo-French Compu-Graphics and Planning Project. He had been sent as an independent auditor to evaluate our project, with the intention to close it. Instead he concluded his mission by writing that our project was very innovative, should continue and the budget should be doubled. He was now working as a consultant to SNCF (the French Railways). And SNCF had a problem that they did not know how to solve. The government had opened the electricity market for industrial users in 2000. SNCF needed to forecast their electricity consumption for every hour of the entire year. They had tried during the first three quarters of 1999 but failed to get satisfactory results. My husband was taken aback as SNCF is full of high level engineers, and replied to Jean-François that he did not know how he could do better than a group of X-Pont engineers. But he gave me the phone and I told Jean-François “if you think they have a need that they are not able to fulfil and you think that the methods that we have developed and you have seen in Hyderabad are useful we can try”. This was the beginning of a seventeen-year collaboration with SNCF as the forecast accuracy that we obtained was excellent and puzzled the engineers of EDF (the French Electricity Board).

I had put a condition to my work with SNCF: I should be allowed to go to India twice a year for two months each time. Although they did not like it much they agreed as the very good forecast results saved them millions of euros. I started working with the Ministry of Rural Development on a DFID project, Andhra Pradesh Rural Livelihood Project, on watershed development integrating physical and financial components as well as impact analysis. Building up of periodical dashboards, with a participatory approach, for stakeholders and decision makers at sub-district, district and state level. I started working with Sanjay Gupta, an IFS officer, posted in Kurnool as the Watershed Programme Director. This was the beginning of a long collaboration, where we started developing visualisation for participatory development programmes. We worked for two main groups, the women Self-Help Groups (SHG), and the Water User Associations (WUA). This led to the development of visual posters for communicating with the stakeholders and giving them tools for comparison with other groups to know where they stand.

In 2005, the IAS officer in charge of the programme, Mr S.P. Tucker, was then posted as Principal Secretary Irrigation, and asked me to develop a GMIS for Andhra Pradesh Irrigation and Command Area Development. As he had been posted earlier in the driest district of the state, Anantapuram, he thought that focussing on bringing irrigation where it was missing, should be the policy that would correct the geographical imbalances. We analysed water related indicators and prepared diagnosis with related risks and potentials. The analysis showed results quite different from the pre-conceived and intuitive ideas of the Secretary. In particular, the Telangana region, which has a relatively high level of irrigation, draws most of its water from underground, with large over-exploitation, indicating unsustainability. It was not sufficient to look at where and how much water, but from which source and for what crop. This led to the creation of a commission to redefine the water policy of the state to include water management, efficiency and sustainability[5]. Looking at potentials we realised that most of the areas with high potential where located in the Tribal Areas. A special Tribal plan was prepared to create new irrigation tanks, and renovate those which had been damaged. All the above material became the base for a water atlas, “Water resources of Andhra Pradesh”, jointly published by IWMI-Tata Water Policy Program (ITP), International Water Management Institute (IWMI) and Visual Information Systems for Action (VIStA), released in 2011[6]. It became the model for the national India Water Resources Information System (WRIS).[7]

In 2008 I was contacted by Mrs Chaya Ratan, IAS, Principal Secretary Women and Child Welfare, to develop a localised (Telugu) GMIS for the Integrated Child Development Services (ICDS), Dept. of Women & Child Welfare, Govt of Andhra Pradesh.[8] The paradigm shift was to place the Anganwadi worker (AWW), the local woman in charge of village women and child welfare, at the centre of the system. Instead of building a traditional monitoring system, that would collect data from the AWW for the benefit of their hierarchy without any useful feedback for her, it was decided to develop a management tool that would provide her with her monthly prioritised task. The first challenge was to develop a computer system in the local language, Telugu, with 52 characters and many combinations, requiring sometime three simultaneous clicks. Instead of using the keyboard, we developed a monitor visual input system, based on the primary school alphabet.

Telugu alphabet displayed on the monitor

The second challenge was to assume that the AWW, most of them with a level of education below 10th class, could use a computer without difficulties. The software was developed to look like the registers habitually filled by the AWW. We developed a training programme based on playful exercises for small children to brake their apprehension. At the end of the first day of training the AWW was able to enter the list of children and mothers. The visual tables that came as output, were immediately understood by the AWW and became a support material for the supervisors. The pilot project was identified as a potential good practice by the central government[9] and received a national award for e-Governance.[10] Unfortunately, although it had been decided to scaled up the pilot project in the entire state, the Principal Secretary, having a disagreement with her Minister, was shifted to another post. The new Secretary did not take up the scaling up. This is typical of the Indian administration, where there is little continuity in actions undertaken by the predecessor and brought forward by the successor (called by IAS officers themselves “successor-predecessor syndrome”!).

In 2012, I started worked for the Andhra Pradesh State Development and Planning Society (APSDPS), Vision 2029 Management Unit (VMU), Planning Department, Government of Andhra Pradesh, as analytics support to the Think Tank entire team of about 30 high qualified professionals from renowned institutes (IIM, ISB, TISS, MANAGE).

Anne Chappuis, born in 1948, PhD (Tropical Geography) 1979. She has spent 49 years in India with 5 years of extensive field work and 44 years work with Andhra Pradesh Government. She is among the early pioneers of visual analytics and started visual data analysis in the seventies. She has in France and in India for local bodies and decision makers on development issues and helped diagnose the problem through maps, graphs and visual matrix, identify similarities and proximities, or oppositions, go from diagnosis to identification of risks and opportunities to develop future strategies and plans.

Ed. Pranav Sharma is a science historian and the curator of the History of Indo-French Scientific Partnership Project

References


[1] https://icaci.org/files/documents/ICC_proceedings/ICC2011/Poster%20Presentations%20PDF/POSTERS%20SESSION%202/P-085.pdf

[2] http://classiques.uqac.ca/contemporains/gaudin_thierry/2100_recit_prochain_siecle/2100_recit_prochain_siecle.html

[3]AUX CARTES CITOYENS, LA DÉMOCRATIE PAR LES CARTES, Anne Chappuis, Luc de Golbéry, Laboratoires MTG et LEDRA, Université de Rouen Cybergeo : Revue européenne de géographie – Cybergeo accueille, n°150, 17 novembre 2000 https://journals.openedition.org/cybergeo/5494?file=1

[4] AUX CARTES CITOYENS, LA DÉMOCRATIE PAR LES CARTES, Anne Chappuis, Luc de Golbéry, Laboratoires MTG et LEDRA, Université de Rouen, in CFC (N°201 – Septembre 2009), http://www.lecfc.fr/new/articles/201-article-4.pdf

[5] Andhra Pradesh State Water Policy, 2008, Irrigation & CAD Department, Government of Andhra Pradesh http://indiaenvironmentportal.org.in/files/AndhraPradeshStateWaterPolicy.pdf

[6] Water Resources of Andhra Pradesh, Sanjay Gupta, Anne Chappuis, SP Tucker, IWMI-Tata Water Policy Program (ITP), International Water Management Institute (IWMI) and Visual Information Systems for Action (VIStA), 2011 https://www.indiawaterportal.org/articles/water-resources-andhra-pradesh-atlas-vista-iwmi-and-goap

[7] https://www.indiawris.gov.in/wris/#/ and GIS layers https://www.indiawris.gov.in/wris/#/lulc

[8] http://enseignelasemio.sciencesconf.org/conference/enseignelasemio/pages/RencontreEnseignerLaSemioAChappuis.pdf

[9] https://icds-wcd.nic.in/Best%20Practices.pdf

[10] https://issuu.com/ehealth_magazine/docs/ehealth-jan-2012–26-28–eindia-2011-award-winners

Geographical research in Andhra Pradesh by French Scholars

Jean Gallais initiated French geographic research in united Andhra Pradesh in 1962 through cooperation with Osmania University and the University of Strasbourg where he was a professor of geography. At the end of the agreement, young researchers from the two geography departments would exchange views to undertake work on fundamental themes corresponding to the research activities of the respective laboratories. The cultural and scientific cooperation advisor of the French Embassy supported the project. It was consequently also financed by the French Ministry of Foreign Affairs.

The first team of four researchers arrived in Hyderabad in 1965, while two Andhra researchers left for Strasbourg. The priority was given to what was the least known, the most problematic and also the closest to Jean Gallais’ research: the development of the rural areas on which he was working in sub-Saharan Africa, an area whose climates were similar to those of Andhra Pradesh and had brought rural people to design various irrigation systems. South India was particularly rich with a palette ranging from micro-farming to giant deltaic developments encompassing millions of farmers.

The approach method, already tested and implemented in Africa, consist of studying in-depth samples of villages selected as representative of the various existing regional agricultural systems. Thus were comprehensively analyzed all aspects of the life of some fifteen villages distributed in the largely rural areas of Andhra Pradesh: Telangana (then not a separate state), the deltas of the Krishna and Godavari rivers, perimeters developed by large contemporary irrigation projects, the Arid Rayalaseema pole of southern India, often mountainous forest areas, habitat of tribal populations, etc.

The first team covered the Telangana, in a team of two, each focusing on a village in North Telangana, one representative of the system of “tanks” common to the entire semi-arid zone of the peninsula, the other in contact with an interstate forest area populated by various communities including tribal.

The second team, also comprising four researchers, studied the market gardening villages around the capital Hyderabad on the one hand, the rice-growing areas of the Godavari Delta and the fertile deltaic belt of black soils on the other.

Each study gave rise to very comprehensive reports, most of which took the form of a thesis. It led to a publication: Villages of Central India.

One of them had, on the occasion of the drafting of his first report, processed their data in the cartography laboratory of the Ecole Pratique des Hautes Etudes directed by Jacques Bertin, founder of Graphic Semiology. The basis of cartographic science includes Graphic Information Processing, the visual equivalent of multivariate statistical processing, and complementary to the latter. He was recruited by the geography department of the University of Rouen-Haute-Normandie to teach this new discipline and the geography of development in India.

Graphic Semiology quickly gained emulators among Rouen teacher-researchers, and a team formed which launched into the computerization of graphic methods on what was the ancestor of the PC, which did not yet exist: the scientific calculator. Thus was born the IMAGE laboratory (for Info-Graphic and Mathematics Applied to Geography) which brought together geographers, cartographers, mathematicians, statisticians and computer scientists. He quickly distinguished himself by developing a set of interconnected programs which allowed the most diverse and powerful interconnected statistical and graphic processing.

The results of these treatments aroused the interest of the director of the Ministry of Planning of the Government of Andhra Pradesh who in 1977 had asked if cooperation between the University of Rouen and his ministry to install an Info- treatment center and possibly train the staff who would operate it to make the most of the information collected by the Bureau of Economics and Statistics (BES). The Cultural and Scientific and Technical Cooperation Advisor of the French Embassy supported the proposal. The program was accepted by the Ministry of Foreign Affairs and began its operations in 1979 under the name of Indo-French Compu-Graphics and Planning Project (CGP in short). It took off in 1980 when a team of French experts was sent to Hyderabad and set up the treatment laboratory at BES.

The program lasted six years during which French researchers trained more than twenty members of the BES, either selected from the staff or recruited from the statistics departments of Andhra universities. The members of IMAGE succeeded each other and made the processing software evolve into a more powerful system, more integrated and easy to use by non-researchers and better adapted to the processing of the massive information that the Office collected to help better planning of the entire state of Andhra Pradesh.

During a visit from J. Bertin, interested in the progress that the program was making in the automation of its methods, he surprised us with his conclusions: “we are now forty years ahead of the minds of potential users of our tools ”; we assured him that the adoption of these innovations was underway. We wanted to see proof of that in the Andhra government’s demands.

The most demanding mission entrusted to him took place when the new Chief Minister N.T. Rama Rao decided to profoundly modify the administrative network of the state to bring the administrative and development services closer to Andhra citizens. It was the reform of the Mandals which consisted of increasing the network of some 180 existing talukas to that of Mandals, which numbered nearly 800. The crucial problem was to choose the capitals of these future entities which had to be equipped with the infrastructure of services lacking to the population. The stakes were considerable both technically (problems of centrality, accessibility, existing infrastructure), financial (cost of new infrastructure) and finally political: local politicians wanted that their villages or towns establishment to become the new capitals.

The CGP was responsible for choosing the future chief towns of Mandal and the villages they would serve. In short, this required gathering all the relevant information at the level of each territorial entity (towns and villages) for the whole of the State, i.e. nearly thirty thousand units. Of the 350 data available on each entity, about fifty were kept for analysis. The first results were quantified and mapped, then sent to government officials in Hyderabad, to the collectors of the 23 districts who were responsible for communicating them to local officials for criticism and substantiated suggestions. These came back to us, were analyzed, submitted to government officials in Hyderabad, entered into the data. A new analysis, quantification and mapping were carried out, sent back to the government and the collectors for a second verification. And this until a result accepted by all stakeholders is achieved. Finally, few proposals posed problems since 94% of the chief towns of mandals were adopted, and the limits altered at the margins. The reform was implemented and the results went beyond expectations and were considered a success, to everyone’s satisfaction.

One day, a phone call from the former cultural adviser who became the diplomatic adviser to the Minister of Cooperation in Paris, who said in substance “we really need here a system like the one you implemented in Hyderabad to help the decision for our actions in Africa. Could you come and install one? “.

Thus in April 1985, we found ourselves in rue Monsieur duplicating the BES laboratory but with the latest innovations in microcomputing, the landscape having been completely upset by the arrival of the IBM PC. In truth our disappointment was great at the discovery of its capacities much lower than the HP9825, the 160 kb floppy disk apart which represented a certain improvement on the cassettes, and the screen despite its more than modest resolution and its graphic functions barely outlines. The processing, therefore, had to continue on the old equipment while the adaptation – de facto the complete rewrite-modification of the software began in parallel. Thus was born the Synthesis and Rapid Information Unit (CSIR) installed in the minister’s hotel.

The processing of all the information collected in the various departments of the ministry, giving rise to the usual analyzes and syntheses, graphical and cartographic, caused amazement in the services. Our situation in the minister’s office gave us access to all information, reports, etc. And their exhaustive visualization revealed and brought out above all the anomalies previously discreetly buried in the tables of figures. We were asked not to circulate them beyond the directions.

After the change of government in 1986, those in charge continued to develop computer graphics tools, one from the IMAGE group that he joined, the other deciding to launch his own company with the objective, beyond the adaptation of software with constantly evolving IT tools, their application in decision support in French organizations of all kinds.

Luc de Golbery is a retired associate professor at VISTA (Visual Information Systems for Action)

Pranav Sharma is a Science Historian and the in-charge of the History of Indo-French Scientific Partnership Project.

Les Recherches Geographiques Francaises En Andhra Pradesh

 C’est Jean Gallais qui amorça la recherche géographique française dans l’Andhra Pradesh unie de 1962 à travers une coopération avec Osmania University et  l’université de Strasbourg où il était professeur de géographie. Au terme de l’accord de jeunes chercheurs des deux départements de géographie s’échangeraient pour entreprendre des travaux sur des thèmes fondamentaux correspondant aux activités de recherche des laboratoires respectifs. Le conseiller culturel et de coopération scientifique de l’Ambassade de France soutint le projet qui en conséquence fut financé par le Ministère français des Affaires Étrangères.

La première équipe de quatre chercheurs, arriva à Hyderabad en 1965 tandis que deux chercheurs Andhra partirent à Strasbourg. La priorité fut donnée à ce qui était le moins bien connu, le plus problématique et aussi le plus proche des recherches de Jean Gallais : le développement des zones rurales sur lesquelles il travaillait en Afrique sub-saharienne, zone dont les climats  se rapprochaient de ceux d’Andhra Pradesh et avaient amené les ruraux à concevoir divers systèmes d’irrigation. L’Inde méridionale en était particulièrement riche avec une palette allant de la micro-exploitation agricole aux aménagements deltaïques géants englobant des millions d’agriculteurs.

La méthode d’approche, déjà testée et mise en œuvre en Afrique, consiste à étudier en profondeur un échantillon de villages sélectionnés comme représentatifs des divers systèmes agricoles régionaux existants. Ainsi furent analysés de manière exhaustive tous les aspects de la vie d’une quinzaine de  villages distribués dans toutes les grands ensembles ruraux d’Andhra Pradesh : Le Telangana (devenu depuis lors un état séparé), les deltas des fleuves Krishna et Godavari, périmètres aménagés des grands projets d’irrigation contemporains, le Rayalaseema pôle aride de l’Inde méridionale, les zones forestières souvent montagneuses, habitat des populations tribales etc.

La première équipe couvrit le Telangana, par équipes de deux, chacune se concentrant sur un village du Nord Telangana l’un représentatif du système des “tanks“ commun à toute la zone semi-aride de la péninsule, l’autre au contact d’une zone forestière inter-étatique peuplée de communautés variées dont des tribaux.

La seconde équipe comptant également quatre chercheurs étudia les villages maraîchers autour  de la capitale Hyderabad d’une part, les zones rizicoles du delta de la Godavari et la ceinture fertile deltaïque des sols noirs d’autre part.

Chaque étude donna lieu à des rapports très complets, prenant pour la majorité la forme d’une thèse. L’ensemble fût résumé en une publication : Villages d’Inde Centrale.

L’un d’eux avait, à l’occasion de la rédaction de son premier rapport, traité ses données au laboratoire de cartographie de l’Ecole Pratique des Hautes Etudes dirigé par Jacques Bertin, fondateur de la Sémiologie Graphique base de la science cartographique moderne qui inclut le  Traitement Graphique de l’Information, équivalent visuel des traitements statistiques multivariés et complémentaire de ces derniers. Il fut recruté par le département de géographie de l’Université de Rouen-Haute-Normandie à la fois pour enseigner cette nouvelle discipline et la géographie du développement en Inde. 

La Sémiologie Graphique fit rapidement des émules parmi les enseignants-chercheurs rouennais et une équipe se format qui se lança dans l’informatisation des méthodes graphiques sur ce qui était l’ancêtre du PC, non encore existant : le calculateur scientifique. C’est ainsi que naquit le laboratoire IMAGE (pour Info-Graphique et Mathématiques Appliquées à la Géographie) qui réunissait géographes, cartographes, mathématiciens, statisticiens et informaticiens. Il se distingua rapidement en développant un ensemble de programmes inter-connectés qui permettaient les traitements statistiques et graphiques interconnectés les plus divers et les plus puissants. 

Les résultats de ces traitements soulevèrent l’intérêt du directeur du Ministère de la Planification du Gouvernement d’Andhra Pradesh qui demanda en 1977 si une coopération entre l’université de Rouen et son ministère ayant pour objectif d’installer un centre de traitement  Info-Graphique et de former le personnel qui le ferait fonctionner dans le but de tirer un parti maximum des informations collectées par le Bureau of Economics and Statistics (BES)  serait possible. Le Conseiller Culturel et de coopération scientifique et technique de l’ambassade de France apporta son appui à la proposition. Le programme fut accepté par le Ministère des Affaires Etrangères et débuta en 1979 sous le nom de Indo-French Compu-Graphics and Planning Project (CGP en bref). Il prit véritablement son envol en 1980 lorsqu’une équipe d’experts français fut déléguée à Hyderabad et installa le laboratoire de traitement au BES.

Le programme dura six ans durant lesquels les chercheurs français formèrent plus de vingt membres du BES soit sélectionnés parmi le personnel soit recrutés dans les départements de statistiques des universités Andhra. Les membres d’IMAGE se succédèrent et firent évoluer les logiciels de traitement vers un système plus puissant, plus intégré et facile d’emploi par des non chercheurs et mieux adaptés aux traitements des informations massives que recueillaient le Bureau pour aider à une meilleure planification de l’ensemble de l’Etat d’Andhra Pradesh.

Lors d’une visite de J. Bertin, intéressé par les avancées que le programme réalisait  dans l’automatisation de ses méthodes celui-ci nous surprit par ses conclusions : « nous avons désormais quarante ans d’avance sur l’esprit des utilisateurs potentiels de nos outils » ; nous lui assurâmes que non et que l’adoption de ces innovations était en cours. Nous voulions en voir la preuve dans le fait les demandes du gouvernement Andhra. 

La mission la plus exigeante qui lui fut confiée eut lieu lorsque le nouveau Chief Minister N.T. RamaRao décida de modifier profondément le maillage administratif de l’état pour rapprocher les services administratifs et de développement des citoyens Andhra. Ce fut la réforme des Mandals qui consista à faire passer le maillage des quelque 180 taluks existants à celui de Mandals au nombre de près de 800. Le problème crucial était de choisir les chefs-lieux de ces futures entités qui devaient être équipés des infrastructures de services manquant à la population. L’enjeu était considérable tant au plan technique (problèmes de centralité, d’accessibilité, d’infrastructures existantes) que financier (coût des nouvelles infrastructures) et finalement politique : les politiciens locaux voulaient pour des raisons évidentes que leurs propres villages ou villes d’implantation deviennent les nouveaux chefs-lieux.

Le CGP eut la responsabilité de choisir les futurs chef-lieux de Mandal et les villages qu’ils dé-serviraient. En bref cela nécessitait de réunir toutes les informations pertinentes au niveau de chaque entité territoriale (villes et villages) pour l’ensemble de l’Etat, soit près de trente mille unités. Sur les 350 données disponibles sur chaque entité une cinquantaine furent conservées pour l’analyse. Les premiers résultats furent quantifiés et cartographiés, puis envoyés aux responsables du gouvernement à Hyderabad, aux collectors des 23 districts qui avaient la responsabilité de les communiquer aux responsables locaux pour critique et suggestions argumentées. Celles-ci nous revenaient, étaient analysées, soumises aux responsables du gouvernement à Hyderabad, introduites dans les données. Une nouvelle analyse, quantification et cartographie étaient réalisées, renvoyées au gouvernement et aux collectors pour une seconde vérification. Et ce jusqu’à ce qu’un résultat accepté par toutes les parties prenantes soit atteint. Finalement peu de propositions posèrent  problèmes puisque 94 % des chefs-lieux de mandals furent adoptés, et les limites retouchées à la marge. La réforme fut appliquée et les résultats allèrent au-delà des espérances et fut considérée comme un succès, à la satisfaction de tous.

Un jour, un appel téléphonique de l’ancien conseiller culturel devenu conseiller diplomatique du Ministre de la Coopération à Paris, qui dît en substance « nous avons vraiment besoin ici d’un système comme celui que vous avez mis en œuvre à Hyderabad comme aide à la décision pour nos actions en Afrique. Pourriez-vous venir en installer un ? ».

 C’est ainsi qu’en avril 1985 nous nous retrouvâmes rue Monsieur dupliquant le laboratoire du BES mais avec les dernières innovations de la micro-informatique le paysage ayant été totalement bouleversé par l’arrivée de l’IBM  PC. En vérité notre déception fut grande à la découverte de ses capacités très inférieures au HP9825, la disquette de 160 ko mise à part qui représentait une  certaine amélioration sur les cassettes, et l’écran malgré sa plus que modeste résolution et ses fonctions graphiques à peine ébauchées. Les traitements durent donc continuer sur les anciens équipements tandis que l’adaptation -de facto la réécriture-modification complète des logiciels débuta en parallèle. Ainsi naquit la Cellule de Synthèse et d’Information Rapide (CSIR) installée dans l’hôtel du ministre.

Le traitement de toutes les informations  récupérées dans les diverses directions du ministère, donnant lieu aux analyses et synthèses habituelles, graphiques et cartographiques, jetèrent la stupeur dans les services. Notre situation au cabinet du ministre nous ouvrait l’accès à toutes les informations, rapports etc. Et leur visualisation exhaustive faisait apparaître et ressortir avant tout les anomalies auparavant discrètement enfouies dans les tableaux de chiffres. On nous demanda de ne pas les faire circuler au-delà des directions.

Après le changement de gouvernement de 1986, les responsables continuèrent à faire évoluer les outils infographiques, l’un depuis le groupe IMAGE qu’il regagna, l’autre décidant de lancer sa propre société avec pour objectif, au delà de l’adaptation des logiciels aux outils informatiques en constante évolution, leur application dans l’aide à la décision dans les organismes français de tous ordres.

Luc de Golbery est professeur associé retraité chez VISTA (Visual Information Systems for Action)