Context
The  agricultural  sector  faces  numerous  challenges.  Simultaneously,  the  development  of digital  technologies  enables  the  generation  of  large  volumes  of  spatialized  data,  which present  a unique  opportunity  to  better  monitor  crops  and  implement  more  efficient,sustainable, and also much more technical agricultural practices. Farmers alone will not cope with  the  scale  and  complexity  of  these  transformations.  More  than  ever,  they  will  need personalized  support  from  agricultural  technical  advisors,  whose  tools  and  methods  must evolve.

UMR ITAP, specifically the DéMo team (Decision and Modeling for Agriculture), and Fruition Sciences  (a viticulture  consulting  company) are partnering within a  LabCom  (ANR  funding).Their joint research aims at developing tools to deploy a 3.0 version of agricultural advisory services:  data-augmented  advisory.  This  involves  integrating  data  acquired  from  the  many and  diverse  sources  of  available  spatial  information,  analyzing  these  shared  datasets  to improve and standardize global expertise, and integrating the operational characteristics of farms to produce locally relevant advice.

We are recruiting a post-doctoral researcher who will participate in the initial work launched by the LabCom starting October 1, 2025 for 18 months.

Job description

State of the art
Agricultural  data,  derived  from  the  technical  and  economic  traceability  of  farms,  are necessarily less structured than data acquired in a controlled experimental context: they are uncertain,  heterogeneous,  asynchronous,  incomplete,  sparse,  etc.  Various  studies  have shown  the  potential  of  this  data  to  generate  new  knowledge  adapted  to  the  local  context(Lamour  et  al.,  2021;  Laurent  et  al.,  2022).  However,  these studies  also  highlighted  the challenges  related  to  the  valorization  of  this  data,  such  as  the  detection  of  local  or  global outliers  (Leroux  et  al.,  2018),  spatial  and  temporal  interpolation  to  address  data incompleteness (Velez et al., 2022), and uncertainty management (Laurent et al., 2022).

Scientific objectives
Data  analysis  approaches  are  constantly  evolving  in  data  science  but  are  relatively underapplied to agricultural data. The first objective of the post-doctoral position is therefore to  conduct a  state-of-the-art  review  of  different  data  analysis  approaches  in  agriculture,highlighting  their  specificities  in  relation  to  the  data.  This  review  can  be  built  from  the example of the databases provided by the DéMo team and Fruition Sciences. The aim is to create a typology of the data to be cross-referenced: multivariate, spatialized (homotopic or heterotopic),  temporal  (synchronous  or  asynchronous),  for  what  purpose  (exploratory,estimation, etc.), and to compare the associated analysis or processing methods, or even to propose a new approach if no method is identified.

By  leveraging  this  new  framework,  the  second  scientific  objective  of  the  post-doctoral position is to implement and evaluate the selected method(s) on example databases, starting with the Fruition Sciences database.

Operational objective
The objective of the post-doctoral position is to develop a method to standardize (merge) an agricultural database, similar to the Fruition Sciences database, to serve two applications: i)an  expert  and ad  hoc  data  analysis  using  a  data  visualization  application,  and  ii)  the implementation of statistical models for the estimation of agronomic parameters.

During  the  last  semester,  the  post-doctoral  researcher  will  work  in  collaboration  with  a computer  developer  from  Fruition  Sciences  to  enable  the  implementation  of  the standardization method in the company's platform and, if time allows, the data visualization application.

Work environment and conditions
The post-doctoral researcher will benefit from co-supervision by researchers from UMR ITAP and  researchers from  UMR  MISTEA,  in  close  relationship  with  Fruition  Sciences,  which  will provide its database and expertise.
The post-doctoral researcher will be hosted at Fruition Sciences premises for part of the week and  at  the  DéMo  team  premises  for  the  other  part  of  the  week,  both  locations  being  in Montpellier, France.

References
Lamour, J., Le Moguédec, G., Naud, O., Lechaudel, M., Taylor, J., Tisseyre, B., 2021.
valuating the drivers of banana flowering cycle duration using a stochastic model and  on  farm  production  data.  Precision  Agric  22,  873–896.
https://doi.org/10.1007/s11119-020-09762-y

Laurent, C., Le Moguédec, G., Taylor, J., Scholasch, T., Tisseyre, B., Metay, A., 2022.
ocal influence of climate on grapevine: an analytical process involving a functional and  Bayesian  exploration  of  farm  data  time  series  synchronised  with  an  eGDD thermal  index:  This  article  is  published  in  cooperation  with  Terclim  2022  (XIVth International  Terroir  Congress  and  2nd  ClimWine  Symposium),  3-8  July  2022,Bordeaux,  France.  OENO  One  56,  301–317. https://doi.org/10.20870/oeno-one.2022.56.2.5443

Leroux,  Corentin,  Jones,  H.,  Clenet,  A.,  Dreux,  B.,  Becu,  M.,  Tisseyre,  B.,  2018.  A general  method  to  filter  out  defective  spatial  observations  from  yield  mapping datasets. Precision Agric 19, 789–808. https://doi.org/10.1007/s11119-017-9555-0 Vélez, S., Rançon, F., Barajas, E., Brunel, G., Rubio, J.A., Tisseyre, B., 2022. Potential of functional analysis applied to Sentinel-2 time-series to assess relevant agronomic parameters at  the  within-field  level  in  viticulture.  Computers  and  Electronics  in Agriculture 194, 106726. https://doi.org/10.1016/j.compag.2022.106726