Our workgroup focuses on sustainable landscape planning and resource management using geospatial analysis, machine learning, and modeling. Lead by @evelynuuemaa
We presented the initial Estonian Environmental Data Cube portal (https://t.co/r8FlE4tCwZ ) at the Estonian Spatial Planning Conference! This portal, developed under #FutureScapes & #WaterSmartLand (funded by #ETAG π¬), offers structured environmental data π for spatial planning
New study alert!π¨ We developed 5 Random Forest models with spatial strategies to predict soil organic carbon ππ±Our top performer? Random Forest Spatial Interpolation for capturing spatial structure. Small but meaningful accuracy gains & spatial covariates often replaced others
We were really happy to be present at the annual Delta Career Day yesterday! π
We met with so many eager students and also had great conversations with them! πβ¨
It was great to see such high interest from the participants towards us! π
#geoinformatics#unitartudelta
Our latest research has been published in Environmental Monitoring and Assessment from @SpringerNature
We analyzed how land cover influences air quality in southern Ecuador using Sentineldata. Thanks to my co-authors @evelynuuemaa,Szilard and Danilo
Read: https://t.co/PMF2Z2ASn7
Last week, our team went on a writing retreat to focus on crafting scientific articles. πβ¨
The days were filled with deep work and collaboration. We also shared valuable tips and tricks on scientific writing tools and methods.π‘
Also, we enjoyed quality time as a team! π₯
We have just opened an exciting PhD position in GeoAI & LLMs and are looking for motivated PhD candidates! The PhD will be supervised by me and @NicoWeghe from the GeoAI Research Center, Ghent University.Β
Apply by March 2nd, 2025 from here (+ details): https://t.co/GRi8MVKxvI
π We're building a #datacube@LGeoinformatics to integrate spatial data with varying resolutions.
Resampling is key for aligning datasets with different resolutions, coordinate systems, or extents.
Check out our quick cheat sheet for resampling tips!
Have any suggestions?π¬
π Last week, @M_Jemeljanova, @allixender & @evelynuuemaa visited @AaltoGeoinfo & @fgi_nls's Geoinformatics Dept. π€
Excited about collaboration between Digital Waters Flagship & ERC #WaterSmartLand, and sharing insights on teaching #FOSS4G.
Thanks a lot for having us! π
π In our @ERC_Research funded WaterSmartLand (https://t.co/8KLZXZ5PUB) project, weβre building an environmental datacube using DGGS tech.
On 17 Dec, @allixender visited #ESA HQ in Paris π«π· to discuss DGGS and share insights with science and technology staff. π€
π In our @ERC_Research funded WaterSmartLand (https://t.co/8KLZXZ5PUB) project, weβre building an environmental datacube using DGGS tech.
On 17 Dec, @allixender visited #ESA HQ in Paris π«π· to discuss DGGS and share insights with science and technology staff. π€
We have Christmas Sweater office day today π€ͺππ§βπlots of cool/funny/weird sweaters and three best were awarded big Grinch Christmas sweater gingerbread π
Last Thursday, we attended the Estonian Remote Sensing Day gathering of experts and enthusiasts exploring remote sensing innovations at @tartuobs ! π°οΈπ‘
π¬ Our team member @holgervirro presented on using deep learning to identify drainage ditches from lidar data.
π° Reflections from @evelynuuemaa on how satellite & open data fuel innovation! π
π°οΈ New models from research, like those at our Centre of Excellence, help to unlock full potential of satellite data.
Read more https://t.co/FpY571RtrB
π· SCANPIX
Finishing up my last work trip of the year with a visit to University of Tartu in Estonia. Taught a short course on Applied Remote Sensing with Google #EarthEngine and had technical discussions with the team on scaling spatial analysis.
Last week, we joined the @Hytruck project quarterly meeting in Riga, where experts in economics, transportation and hydrogen discussed the future of hydrogen-powered freight transport!
We got great feedback on our work and explored hydrogen-powered public transport in Riga! π
π§ Key Insights:
β’RF excels with historical flows & time-lags π
β’Shows robust daily performanceπͺ
β’SWAT struggles with daily forecasts, overestimating by 27%! π²
While SWAT shines in monthly predictions, RF is the go-to for daily accuracy, especially in arid regions!ππ‘
ππ Our recent study proposes a novel time-lag-informed Random Forest (RF) framework that outperforms traditional methods for streamflow predictions! π See key insights in the comments below. Paper: https://t.co/EhcvWAfNtR