A Quick Test Drive of Google Table Fusion

Google Labs launched a cool new app yesterday called Google Table Fusion Although, the name is a bit confusing since there is also the Google Earth Fusion Server. Anyways, it is a great step forward into the wide world of structured data for Google. This seems to be becoming a trend following the launch of Google Public Data a little while back. The future of structured data on the Web is probably a whole discussion in itself, as well as what Google’s entrance means, but I’ll keep to just the new app for the sake of brevity.

Google Table allows users to upload up to 100mb .csv or .xls file and then allows filtering, merging and visualization of the data. Each of the basic options has lots of nice functionality. Filter allows you do some basic operators like > “greater than” = “greater than or equal to” < = "less than or equal to" against any single numerical attribute in the table. You can also merge two tables together either in total or by attribute. Lastly, Google Table provides a variety of data visualization options – map, intensity map (i.e. thematic map), line graph, bar chart, pie chart, scatter plot, motion (Gapminder's stuff), and timeline.

With that as a high level overview I'll dig into geographic capabilities a bit, since we tend to like maps. On the plus side this is the first I've seen from Google directly linking structured data with maps. Google Charts has had thematic maps for a while, which is leveraged for the thematic mapping in Google Table. Also Google Table has some georeferencing capabilities that auto-recognizes geographic names and georeferences them on the fly.

As MapperZ points out Google Table has "fast batch geocoding" capabilities- except you can only look at the geocoded files 200 at a time and not get the georeferenced data out of the system. It appears that Google is georeferencing a set on place names and addresses – from what I can gather this includes country names, city names (domestic and international) in addition to doing traditional street address geocoding. The georeferencing and geocoding appears to only work in Google Maps unless it is country names. If you have country names then there is an option to visualize the data in Google Charts. This can cause some issues – for example look at this state GDP data I uploaded. The first pic is data geocoded by clicking on the "Map" option:

The problem becomes even more severe with Google Chart’s georeferencing where it thinks the state of Georgia is the country of Georgia since that is the only country name it recognizes:

These are problems all geocoders run into, but I think it is endemic of where Google Table comes up short. It is a bit of an application Frankenstein where it does many things but none of the very well. Several different Google apps glued together, but without the cohesive work-flow and technology to make it work as a platform. The integration and capabilities for data search, metadata, handling data other than .csv, large data visualization,and ease of use will all be keys for success. Also in the negative category, the big disclaimer, any data you upload into Google Table incurs the dreaded section 11 clause:

By submitting, posting or displaying the content you give Google a perpetual, irrevocable, worldwide, royalty-free, and non-exclusive licence to reproduce, adapt, modify, translate, publish, publicly perform, publicly display and distribute any Content which you submit, post or display on or through, the Services.

Criticisms aside it is pre-alpha and a “Labs” project. There are many cool innovations like commenting within data sets. Also I think Google Table is a trend we’ll see more of where the geographic components of data is just another option in a larger system and not a system in and of itself. The overall project points towards the big potential in building a cohesive structured data platform, and it will be fun to see which of the various efforts in the ecosystem wins the day.