1. Definition and Analysis of New
Agricultural Farm Energetic Indicators
using Spatial OLAP
Sandro Bimonte, Kamal Boulil, Jean-
Pour mieux
affirmer
ses missions,
Pierre Chanet, Marilys Pradel
le Cemagref
devient Irstea
Irstea, Clermont-Ferrand, France
Sandro.bimonte@irstea.fr
www.irstea.fr

2. 2
Plan
Spatial Data Warehouse and Spatial OLAP
Agricultural farm energy consumption diagnosis and related
indicators
New energy indicators to assess farm performance at a detailed
scale
Spatial OLAP analysis of new indicators
Conclusions

3. Spatial Data Warehouse
and Spatial OLAP
Pour mieux
affirmer
ses missions,
le Cemagref
devient Irstea
www.irstea.fr

4. 4
Data warehouse and OLAP
Data warehouse is: "A collection of integrated, subject oriented,
time-variant, and non-volatile data from various sources into a
single and consistent data repository to support decision-
making” (Inmon, 1996)
Hypercube
Temps
Time
4000 1000
Facts, measures, dimensions, and hierarchies 8000
Client
Costumer 7000
Aggregations (SUM, MIN, MAX, AVG, COUNT)
3000
8000
12000 1000
OLAP analysis 2000
6000
Drilling, Slicing, pivot, etc. 8000
Produit
Product
Interactive analysis and aggregation of huge volume of data
•What is the total of sales per month, client and product ?

5. 5
Spatial OLAP (1/4)
Spatial OLAP : "A visual platform built especially to support rapid and
easy spatiotemporal analysis and exploration of data following a
multidimensional approach comprised of aggregation levels
available in cartographic displays as well as in tabular and diagram
displays” (Bédard, 1997)
Cartographic representation:
Visualization of distribution of facts
Visualization spatial relationships between
facts and dimensions
Visualization of facts at different granularities

6. 6
Spatial OLAP (2/4)
Spatial data warehouse: "A collection of integrated, subject
oriented, time-variant, and non-volatile spatial data from
various sources into a single and consistent data repository to
support decision-making” (Stefanovic, et al.2000)
Spatio-multidimensional model:
Spatial dimesion: Dimension with spatial attributes
Spatial measure: Geometry and/or results of spatial operators
Spatial Aggregation (e.g. UNION, INTRESECTION)
SOLAP operators: Drilling in spatial dimension, slicing the hypercube uing spatial
predicats

7. 7
Spatial OLAP (3/4)
 What is the total sold by month and city at 50 Km far from
Paris

8. 8
Spatial OLAP (4/4)
Talend Oracle Spatial, GeoMondrian,
Tools Jrubik, GeWOlap, …
SDI PosteGIS, … Map4Decision,
…

9. Agricultural farm energy
consumption diagnosis and
related indicators
Pour mieux
affirmer
ses missions,
le Cemagref
devient Irstea
www.irstea.fr

10. 10
Agricultural farm energy consumption
diagnosis (1/4)
Main goals:
- Quantify energy consumption per processes to identify possible improvements by
acting either on the production system, practices or equipment
- Compare energy performance of livestock and crop farming
- Establish reference values for the productions
Energy indicators : Variables that provide information on avaiable data, used to
take decisions
simple (e.g. fuel) or aggregated indicators (e.g. energy balance)
Inputs are (i) direct and indirect energy flows and (ii) energy consuming operations
Direct energies (e.g. electricity, fuels and gases)
Indirect energies are energies used to produce outputs (e.g. fertilizers, crop
protection products)

11. 11
Agricultural farm energy consumption
diagnosis (2/4)
Diagnosis are grouped:
Global agro-environmental diagnoses: these consider the farm as a whole and
assess the global environmental impact of the farm
(Sum of the quantities of direct and indirect energy used
by the farm, expressed in MJ per year)
Field agro-environmental diagnoses: these assess the environmental
performance of the farm at the field scale
(Sum of the quantities of direct and indirect energy used
by the farm, expressed in MJ per hectare per year)
Sustainability diagnoses: these assess farm sustainability performance through
environmental, social and economic sustainability issues

12. 12
Agricultural farm energy consumption
diagnosis (3/4)
Energetic diagnoses: these allow an energy balance at the farm scale by
quantifying energy inputs and outputs
Prediagnoses or autodiagnoses: these assess the energy consumption of farm
equipment and give an idea of the main possible improvements
Life cycle assessment: these methods assess the environmental impact on a
system by inventory pollutant emissions, raw materials and energy during its
whole life cycle
(Energy consumption based on the functional unit used,
either the hectare or milk liter)

13. 13
Agricultural farm energy consumption
diagnosis (4/4)
Main limitations:
they define indicators only at a global scale, avoiding the finest analysis of
energy consumption related to technical operations and plots
they are based on classical data storage systems that do not allow
interactive cartographic analysis of huge volumes of data
An analysis at a most detailed scale (field, production activity or operation) will
require:
1. More precise data acquisition systems
2. A set of new indicators
3. Systems allowing visual/cartographic interactive analysis at different
granularities/scales of huge data volumes

14. New energy indicators to
assess farm performance at
a detailed scale
Pour mieux
affirmer
ses missions,
le Cemagref
devient Irstea
www.irstea.fr

15. 15
New energy indicators (1/4)
Analysis scale:
Spatial scale: the hectare is the spatial reference unit used to express the
technical-economic results on a farm
Temporal scale: the hour is the temporal reference unit for farmers
This unit is frequently used to express fuel consumption for mobile equipment
and electric/gas consumption for electric equipment
Production scale: all of the energy flows involved in farm activities (crop or cattle
management) can also be expressed at the production scale
The results will be expressed with the common production unit used by the
farmer depending on the type of farm production (e.g., tons of dry matter, liters
of milk, etc.)

16. 16
New energy indicators (2/4)
These indicators were calculated for the three main farm activities:
Crop management activities: technical operations on crops (sowing, plowing,
fertilization, harvest, etc.)
Cattle management activities: technical operations on cattle, such as care,
feeding and milk/meat production (milking, slaughtering, etc.)
General activities: technical operations that cannot be allocated to cattle or crop
management (cleaning, logistics, transport, etc.)
These indicators can be aggregated at the global scale using the sum

17. 17
New energy indicators (3/4)
Example simple indicator: the fuel consumption per plot, technical
operation, year and production
“140 liters of fuel used for the parcel ‘13 pal’ of the
farm of Montoldre in 2010 during the plowing operation for
the production of wheat”
Example aggregated indicator: fuel consumption per farm, technical
operation, year and production
“The aggregates on spatial scale are calculated by summing
all the plots' values belonging to the same farm”

18. 18
New energy indicators (4/4)

19. Spatial OLAP analysis of
new indicators
Pour mieux
affirmer
ses missions,
le Cemagref
devient Irstea
www.irstea.fr

20. 20
Spatial OLAP analysis of new indicators (1/9)
A SOLAP system to represent, store and analyze the new indicators
→ interactive analysis of huge volumes of (aggregated) indicators values and it
provides a cartographic representation
New indicators and spatio-multidimensional model:
Indicators = measures
Inputs (energy, time, spatial scale, technical operation and production) =
dimensions
Aggregated indicators = aggregated measures

21. 21
Spatial OLAP analysis of new indicators (2/9)
Usage of a UML profile implemented in the CASE tool Magic Draw
Automatic implementation on the SOLAP architecture
→ fast and flexible schema development
Stereotypes for each spatio-multidimensional element
e.g. "Fact" for the facts, "SpatialAggLevel" for spatial dimension levels
A stereotype is an extension of a UML element (class, attribute, etc.=

22. 22
Spatial OLAP analysis of new indicators (3/9)
“Fuel to weed a plot of wheat"
“Total consumption of fuel to weed the entire farm per year"

23. 23
Spatial OLAP analysis of new indicators (4/9)
Dimensions:
Campaigns (Campagne): production cycles expressed in years (e.g., wheat
produced in 2009)
Time (Temps): classical temporal dimension, in which days are grouped by month
and year
Products (Produits): the input and output products (intrant and extrant). Products
are grouped recursively into larger classes of products
Operators (Opérateurs): people who perform the operation
Equipment (Attelage): machines and tools used
Location: the spatial dimension that groups plots (parcelle) by farm (exploitation),
department (département) and region (région)
Productions: type of production (e.g. wheat)
Technical Operations (Opérations Técniques): the technical operations performed,
which are grouped by functions

24. 24
Spatial OLAP analysis of new indicators (5/9)
Measures:
the area worked (surface_w)
the number of animals (animaux_nb)
the amount of product (input represented with “intrant” and output denoted with
“extrant”) used during work or no work (denoted with “w” and “nw”,
respectively)
the duration in hours (duree_w and dure_nw)
the distance traveled (distance_parcourue_w and distance_parcourue_nw)
The measures are aggregated using the sum along all dimensions

25. 25
Spatial OLAP analysis of new indicators (6/9)
Implementation:
SDW tier: PostGIS is a Spatial DBMS natively supporting spatial data
SOLAP Server and Client:JMap-SOLAP only existing SOLAP tool

26. 26
Spatial OLAP analysis of new indicators (7/9)
Aggregated indicator "the amount of fuel used for cultivating one hectare of
wheat on the Montoldre farm”

27. 27
Spatial OLAP analysis of new indicators (8/9)
Spatial drill-down operation: “energy consumption per plot” indicator

28. 28
Spatial OLAP analysis of new indicators (9/9)
Drill-down on the technical operations dimension: “more details on energy
consumption by plot”

29. 29
Spatial OLAP analysis of new indicators
“fuel consumption per plot, technical operation, year and production”

30. 30
Conclusions and perspectives
We propose some new indicators to assess agricultural farm
performance at a most detailed scale
We also show how it is possible to represent these indicators
using a spatial data warehouse and analyze them by means of
a classical SOLAP system
We are working on the estimation/interpolation of missing values
at most detailed dimensions levels