Services
Consulting and technical expertise
- Audit of GIS data engineering practices
- Requirement gathering and writing of technical specifications
- Design of Spatial Data Infrastructures
Development
- Design and development of GIS data engineering pipelines
- Big Data and cloud solutions for large scale data processing
- Complex GIS-oriented algorithms (spatial analysis, Machine Learning)
- Design and development of API's and data / service platforms
- QGIS plugins
Resume
Portfolio
WANAKA
Design & development of a remote sensing data processing pipeline for precision agriculture
Details...
Context
- WANAKA is a SaaS publisher of Decision Support Tools for precision agriculture
- Its flagship product, Agrorendement, provides fertilizer management guidance for wheat, barley and rapeseed: It requires the processing of a large amount of remote sensing data (drone, airplane, satellite) with very tight deadlines
- Originally, the processing was done manually by combining several tools (ArcGIS, Excel, Farm Works)
Solution
- To cope with the ever increasing volume of data, I was commissioned by WANAKA to design a fully automatic processing pipeline for remote sensing data
- Based on agronomic models previously designed by their R&D, I developed an initial prototype with Python, ArcPy and the ArcGIS ModelBuilder
- After validation during a first growing season, I replaced this prototype with a highly optimized processing chain based on Python and open source libraries (RasterIO, NumPy, Shapely, Matplotlib) and deployed on Google Cloud Platform
Result
More than 300,000 ha were processed between 2016 and 2019: The pipeline generates PDF
reports detailing analysis results for all WANAKA customers, as well as variable rate
application files (SHP, RDS, XML ISOBUS formats)
ArchiFiles
Audit of a GIS data production process & development of QGIS plugins
Details...
Context
- ArchiFiles produces indoor GIS data (shopping malls, museums, airports...) on behalf of cartographic portals such as Google Maps and Apple Maps
- Originally, their team of 150 operators used a variety of tools (Google Earth, AutoCAD, ArcGIS) to produce GIS data
Solution
- With the goal of improving productivity and quality, I was commissioned by ArchiFiles to audit their data production process, analyze the possibility of consolidating production on a single software and identify automation opportunities
- I recommended migrating the entire data production to QGIS and wrote specifications of QGIS plugins for additional features to develop
- In a second phase, I developed the plugins, together with a team of 2 programmers internal to Archifiles whom I trained on Python and QGIS
Result
More than 3,000 indoor venues were digitized in 2014 and 2015 with the production process
based on QGIS
Total
Design of a GIS data processing pipeline & development of a cartographic portal
Details...
Context
- Total Marketing & Services (M&S) manages the distribution and storage of oil-based products in France (gas stations, oil terminals, refineries)
- A Spatial Data Infrastructure had been created in order to support crisis management. However, out-of-date data, caused by a long and unreliable data integration process, coupled with bad usability of the desktop interface used for map visualization, resulted in poor adoption
Solution
- I was commissioned by the HSE division of Total M&S to make the acquisition and processing of data more reliable, as well as find ways to extend their use within the company
- I designed and developed processing pipelines with Python, ArcPy, and FME in order to 1) integrate internal and external GIS data into a common database and 2) automate the creation and deployment of maps
- I created a metadata catalog detailing provenance, schemas and processing for all the GIS data
- I proposed and developed a portal and an ArcGIS Server-based web interface for easy access to maps
Result
The time required to integrate the data went from one full-time engineer previously to a
few days per month by a technician. The new web-based data visualization interface ended up
being used by dozens of users throughout Total M&S