Giles Lane came out with a new initiative to build on top of the urban tapestries platform. The idea is to build a wearable biosensor kit that should interface with a mobile telephone and transfer rhythmically some parameters to the urban tapestries system so that the parameters can be mapped to a certain part of the city. Basically the idea is to use the mobile as a remote and distributed sensor to gather environmental/personal data.

Sensory Threads aims to stimulate and inform a public debate on how personal biosensor data is collected and used in biomedical science. This project will combine an artistic with an evidence-based approach, building and testing a prototype body biosensor network that uploads data to Proboscis’ Urban Tapestries spatial annotation system. This will allow participants to map experiential and emotional annotations to their readings – adding a whole new layer of sentient knowledge to machine data. It is intended to demystify how data is collected, what it produces and how we can correlate it to other factors affecting health such as environmental pollution.

Connected pasta: I wrote already of this concept here.

UPDATE from Giles: What Sensory Threads will focus on is the ethics of using (and developing) these kinds of systems… how will it affect us socially and culturally — what are the implications of correlating the machine data of the sensors with the human/sentient knowledge of place/context/situation etc…

Tags: ecology, environment, Exploratory Data Analysis, field tools, hack, maps, psychogeography, urban exploration

Computers make it easier to do a lot of things, but most of the things they make it easier to do don’t need to be done.

— Andy Rooney

SciPy_Core is a module for python that replace the old Numeric. It has lots of great features and improvements. I started to use it because I was in the need of some complex matrices manipulations that are just a pain (or impossible) to do with lists.

Travis Oliphant is currently maintaining the project and the documentation.

Tags: Python

1) So far the CLARANS algorithm (Clustering for Large Applications RANdomised Search) seems to be the most promising because of the implementation that specifically targets geographical applications. Additionally, two variants of this algorithm seems to allow to give more or less prevalence to the semantic component of my specific case:

SD(CLARANS) are the spatial-dominant variant and NSD(CLARANS) is the non-spatial-dominant.

2) While cluster analysis sometimes uses the data matrix, some clustering algorithm uses the similarity matrix, that is often called proximity matrix.

In my case I have two proximity matrices, the first is generated by the two-dimensional geographic distance of the messages. The second is the semantic distance as returned from a comparative method based on Latent Semantic Analysis, for instance.

One of the idea is to use these two matrices alternatively on the algorithm or to combine them before passing the average to the cluster engine.

3)  An interesting technique called Sparsification allows to reduce the connectivity of nodes by breaking some of the links of the connectivity graph.  This is really interesting and can be used to take into account links from one message to the other and an external data enrichment to weight the incidence of specific nodes.

Tags: clustering, map algorithms, spatial clustering, statistics

I was baking a fresh piece of bread on this augmented Kitchen a couple of years ago with my friend Jofish. It was cool to see how new technologies could support human activities to make cooking experience more fun.

However nowadays I would be really more amazed to see an attitude to extend these concepts to something that isn’t just fun or cool. I would love to see some projects for energy savings, material recycling, etc., lie in the examples proposed by Beppe.

Tags: information visualization, interaction design, interactive furniture, new technology, politics

Jeff Israely wrote an article on Time about one of my preferred italians: Beppe Grillo. He was mentioned as one of the 2005 European Heroes. In the article, the author makes a review of his career, telling some bits of his life and what brought him to target politics.

Maybe Beppe Grillo should have been a state auditor instead of a comedian. Back in December 2003, Italy was rocked by the Parmalat scandal, in which the Italian dairy conglomerate imploded in a vacuum of debt and deception that became the largest bankruptcy in European history. But Grillo had seen it coming.

I do not laugh any more to the Italian situation. Sometimes I even thought that I did not want to be Italian if being Italian was being part of this current situation. I really think that we are loosing those traits that made us so cool, like our values: the family, an honest work, our creativity. Grillo is  one of those few that makes me proud of being Italian. Again, is one of those few that raise the head from the mood and says: Hey, wait a minute!

Sometimes, when body politic is sick, laughter is the best medicine.

Tags: politics, Italy, society

Today I discussed briefly the algorithm for my thesis with my advisor. We decided to not spend to much time trying to support multi-languages analysis in the system. Then we decided also that a better way to go is to use semantic distance an heuristic for spatial clustering. This basically brings me to a choice:

1. I find a way to translate a matrix of semantic distances into some form of values to pair with the geographical two-dimensional space. In this case I can proceed to checkout with an available clustering method with non-geographcal features.

2. I find a way to “parametrize” the clustering using the semantic distances… not clear how.

3. Compute with the semantic distances a semantic coherence that is the output of the spatial cluster. Then running many times the spatial clustering to maximize the outcome of this coherence.

Tags: map algorithms, spatial clustering

H. J. Miller and J. Han. Geographic Data Mining and Knowledge Discovery, chapter Geographic data mining and knowledge discovery: An overview, pages 3–32. Taylor and Francis, London, 2001.

————————-

This book introduction presents an overview of the methods for kwoweldge discovery on geographical systems. It ferst present an overview of the techniques from the pre-processing with data enrichment  to data-reduction and projection and finally data mining.

The data mining techniques are divided into segmentation (clustering and classification), dependence analysis, deviation and outlier analysis, trend detection and generalization. For each phase some example and relative literature is presented.

Spatial Clustering can be based on a combinations of non-spatial attributes, spatial proximity, time, ad spatial attributes. Research on the computer science side focused on several salable algorithms for clustering very large spatial datasets and methods for finding proximity relationships between clusters and spatial features (Knorr and NG, 1996). 

Tags: clustering, spatial clustering, statistics

I spent some time today trying to figure out which clustering packages are already available for python. Seems that all the bleeding-edge clustering methods like CHAMELEON, BIRCH, CURE, CLARANS are still missing an available implementation. Actually these methods are missing much more than a python porting. They are listed in the literature but not available to the wide public. I wonder if this is just another copyright issue …

However, I found a nice package for C that is called, guess what?: Cluster. This package implements the k-means, k-medians, k-medoids, treecluster, Self-Organising Maps. Not much but better than nothing. A Python meodule is also available: Pycluster.

A more complete environment seems to be this gCLUTO (CLUstering TOolkit). However doesn’t seems to be easily callable from Python. I have to enquiry more.

Tags: algorithmic information theory, clustering, spatial clustering, statistics

This is the theme of the presentation of Fabio Sergio at MEX 2005 (Mobile User Experience) – September 6th, London.

“” (2.5 Mb, PDF)

Synopsis:

As we look around us inanimate things are slowly coming to life, veneered with layers of digital information. From payments made by touching things to street signs that broadcast messages our belongings are moving from supporting our behaviors to developing their own.

In the brave new world mobile connected devices will be at the center of the convergence of wide-band wireless connectivity, RFID and (A)GPS-enabled applications. They will stop being purely at the receiving end of data streams and become conduits, mustering bits from objects and infusing them into other objects.

How will all of this impact the design of mobile-mediated experiences?

Are we moving towards a world of seamless socio-economical transactions or rather towards a permission-based reality, plagued by constant confirm/cancel requests?

What new scenarios will be driven by these innovations?

Tags: CellID, Location Based Services, new technology, usability