Building a data agency in Belgium from scratch. -data strategy, analytics, viz- Honorary research fellow at @CASAUCL Previously at @OpenLab_Ncl, @NYU_CUSP
Belgium
Joined May 2012
- Tweets3K
- Following1.9K
- Followers657
- Likes2.6K
Maarten Vanhoof retweeted
This week when I announced the release of my free e-book, “Applied Machine Learning in Python: A Hands-on Guide with Code,” I promised it would be a living document, continuously updated with new content. 📘✨
@GeostatGuy delivers! 🎉🚀 Yesterday, I added a new chapter on density-based clustering. Check it out @
geostatsguy.github.io/Machin… #DataScience 🔍📊💡
Spectral clustering, artificial neural networks, convolutional neural networks are next. Any requests? I’m stoked to help folks learn.
Maarten Vanhoof retweeted
Prof @jmichaelbatty has co-authored an article in the first issue of @NatCities / @NaturePortfolio. It proposes using data from the world’s 7+ billion mobile-phone users as a proxy of the temporal and spatial connections that make cities what they are.
ucl.ac.uk/bartlett/casa/news…
Idem! I'll be attending as well, always happy to have a chat and share some laughter!
Seeing posts from lots of people going to informationplusconference.co… Looking forward to see you all there, and meet old and new friends!! #dataviz
I've got a particular affection for the census travel to work data (perks of my previous job @CASAUCL).
This visualisations is 100% igniting that affection!
Some weekend R&D in visualising some of the census 21 travel to work flow data per MSOA. Each sphere is proportional to volume/10. I like using particle flows in these cases to comprehend flow directionality. #map #3d #animation
Interesting, and beautiful..
It's particularly windy in the north west today as a result of Storm Debi passing through 🌬️⚡
Renewables are currently making up ~60% of generation in Great Britain
See for yourself on the renewables map:
renewables-map.robinhawkes.c…
It's marking season!
Every year I buckle up to mark a number of mastertheses from @CASAUCL.
Here's 4 reasons why I do this.
3. It's a good way to keep tabs on the skills and level of graduates you can expect.
At my own company (a data agency) I see a lot of young applicants.
Having a reference point on what you can expect from them technically is a very useful benefit of reading their work.
This is amazing.
Probably the best explainer of large language models I've seen, ever.
Great job @inari_ta and team!
Our latest visual story explains how Generative AI works, what are transformers and why “large language models” are such incredible tools.
Done with this awesome team @Dan_Clark5
@sam_learner @samjoiner @madhumita29 @olihawkins and @EadeMoon!
👉 ig.ft.com/generative-ai/
Michele van marketing draait overuren!
Wil je wel eens weten wat ik net doe als job?
Lees dan zeker ons interview op:
craftzing.com/what-we-think/…
How to erase a night within seconds:
1. You learn that the Information is beautiful awards submissions are in.
2. You navigate to: informationisbeautifulawards…
Thank me later.
Ik vraag me af of @StynenCharlotte door heeft dat de jongeman die gisteren zijn debuut voor @UnionStGilloise maakte opgegroeid is in dezelfde straat als zij.. 😍
Honestly don't know what to think about this..
Especially since Folsom Prison Blues has always been a favorite of mine.
Like Johnny said: I'll be damned.
Maarten Vanhoof retweeted
Onder leiding van Maarten Vanhoof wil het bedrijf zijn expertisecentrum rond data, AI en machine learning op korte termijn uitbreiden tot dertig specialisten. datanews.knack.be/carriere/c…
Maarten Vanhoof retweeted
📢 Our JRC Science for Policy Report "European Data Spaces - Scientific Insights into Data Sharing and Utilisation at Scale" is out‼️
It offers our perspectives & lessons learnt on data sharing to inform the establishment of 🇪🇺 data spaces.
Access it at europa.eu/!RBQXmx
If you're looking for an interesting data challenge: look no further
The NetMob 2023 Data Challenge was launched today. Don't miss the opportunity to access a unique mobile network dataset. Apply now at netmob2023challenge.networks… !
Dataset details here: arxiv.org/abs/2305.06933
Interesting, especially the community detection part. It's a logical idea but not that easy to implement en maintain in my opinion..
Twitter has released the code for its recommender. I am reading through this documentation and the actual repo and will have commentary soon. blog.twitter.com/engineering…