Big data meets scalable visualizations JAVIER DE LA TORRE
Dec 24, 2014
Big data meets scalable visualizations
JAVIER DE LA TORRE
���2Javier de la Torre - @jatorre
Big data meets scalable visualizations
�3
picture on big data awesomeness
Big data awesomeness!!!!
�4
�5
�6
Big data without data visualization
= #fail
Maps are the most popular type of data visualization
Everything happens somewhere !Where are your clients? IP=location !So everything can be analyzed and visualized on maps
Everybody wants to see data on maps,
But making good maps is very hard!
Ugly map!
Making maps is hard because…
Tools are not there yet. They are for GIS experts !Handling 100 points is easy, 1Million is hard !Data chages! Is not about printing maps online!
�11
Demo on meteorites
Wall Street Journal US election maps
Big data analysis and reporting tool - UNEP Carbon calculator
Narrative maps / Story telling - The Hobbit filming Locations map
Narrative maps / Story telling - The Rolling Stones tour maps
German elections real time maps
Visualizing NYC Open Data
Animated geotemporal maps. Everything happens somewhere and at some time. Navy of WWI map
Visual analysis - Economic impact of the Mobile World Congress 2012 in Barcelona
All meteorites fallen on earth
Animated city traffic maps
Mobile ready.
�23
Big data analysis of deforestationHow we can track deforestation on real time Global Forest Watch
http://en.wikipedia.org/wiki/Bakun_Dam
Most people don’t need Big Data technologies
But when you can’t…. when it really explodes…
You just need to start collecting and analyzing data. Don’t focus on technology, probably your database can already do it !You are not Facebook, don’t be cheat
Foreign Data wrappersConnect PostgreSQL to almost anything
Oracle Hadoop MySQL MongoDB CouchDB Redis …. Twitter Email S3
�49
CartoDB Hadoop HBase
Geo-temporal visualizationsCartoDB and Torque
Friday, 28 September 12
WITH%hgrid%%%%%%AS%(SELECT%Cdb_rectanglegrid(Cdb_xyz_extent(8,%12,%5),%%%%%%%%%%%%%%%%%Cdb_xyz_resolution(5)%*%4,%%%%%%%%%%%%%%%%%%%%%%%%%%%%Cdb_xyz_resolution(5)%*%4)%AS%cell)%SELECT%x,%%%%%%%%y,%%%%%%%%Array_agg(c)%vals,%%%%%%%%Array_agg(d)%dates%FROM%%%(SELECT%St_xmax(hgrid.cell)%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%x,%%%%%%%%%%%%%%%%St_ymax(hgrid.cell)%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%y,%%%%%%%%%%%%%%%%Count(i.cartodb_id)%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%c,%%%%%%%%%%%%%%%%Floor((%Date_part('epoch',%built)%Q%Q10418716800%)%/%32837875)%d%%%%%%%%%FROM%%%hgrid,%%%%%%%%%%%%%%%%us_po_offices%i%%%%%%%%%WHERE%%St_intersects(i.the_geom_webmercator,%hgrid.cell)%%%%%%%%%GROUP%%BY%hgrid.cell,%%%%%%%%%%%%%%%%%%%Floor((%Date_part('epoch',%built)%Q%Q10418716800%)%/%32837875)%%%%%%%%)%f%GROUP%%BY%x,%%%%%%%%%%%y
Friday, 28 September 12
{%%rows:%[%%{%%%%x:%0,%%%%y:%0,%%%%vals:%[2],%%%%dates:%[457]%%},%%{%%%%x:%1,%%%%y:%0,%%%%vals:%[1,1,4],%%%%dates:%[2,3,4]%%%%}%%]}
Friday, 28 September 12
1
10
100
1000
3mb 70mb 300mb1.5
21.2
300
70
3
Raw Datacube
Payload sizes
Friday, 28 September 12
Think on the value of location on your data, and use it!
Is very likely you have geospatial data already !Complete the big data cycle: Don't forget data visualization !Find the stories inside the data and show them!
���56Javier de la Torre - @jatorre
Thanks!!