Have you ever traveled out of the state you live in and found yourself saying, “Wow, people in this state are terrible at driving.” Now you can see if your claim was appropriate after looking at the GMAC Insurance National Drivers Test.

GMAC Insurance has been conducting an annual survey where respondents take a driving test that contains questions from DMV tests across the country. Below is a map of the average scores from 2010 along with their inverse ranking among the 50 states and the District of Columbia.

View full map

From the map you can see that states in the darker orange color range had the highest scores and states with the lighter orange colors scored lower on the test. On the 100 point scoring scale the highest state score was Kansas with 82.3 as their average. The lowest scoring state was New York with a score of 70.0.

There was also a second part to the survey. This part surveyed drivers on the types of distracting behavior that they took part in while driving. These distracting actions include applying makeup, changing clothes, eating, talking on a cell phone, and texting on a cell phone while driving. Below is a map of the percentage of respondents per state that responded to participating in these distracting behaviors.

View full map

The above data is all very interesting and I wondered to myself what might cause the bad driving statistics? I decided to then correlate the average scores from the GMAC Test with three types of data: 1. Max State Speed Limits by State (to see if fast driving correlated to bad driving) 2. % of Deficient Bridges by State (see if poor road conditions correlate to bad driving) 3. Population Density by State (to see if congestion correlates to bad driving). These are not perfect indicators, but I thought it might be fun to see of any of these numbers might correlate strongy. Below are the maps:

View full map

The correlations are interesting:

We see that the max speed limit vs. the average scores had a low correlation of .39. So it is probably safe to say that slow max speed limits or high max speed limits do not deter people from being bad or good drivers. Bridge conditions had a slightly stronger correlation at -.49. This is a bit stronger and may hold some weight for arguments sake. Then the last correlation of population density we see as the strongest at -.56. Also not extremely strong but may be something to consider when deciding why people are bad drivers in certain states.

I found the data from GMAC Insurance to be rather interesting and had fun looking at my state and other states that I have traveled through. See what you think of the results and see if you can see why drivers from Kansas score better than drivers from New Jersey.

Popularity: 10% [?]

It has been great to see a dialog emerge on open data surrounding ESRI’s merging of Geodata.gov and Data.gov. I think Marten’s post title sums up my general consensus – “confusion”.

Every person I’ve talked to and every post on the topic has a different version of the story. Further, the story changes as the community asks hard questions. It is all a bit murky, but I thought this was a good opportunity to take a step back and encapsulate what the community is looking for with open data. In this case I’m going to lump together the geo-community and the broader data community because the two are converging at a rapid pace.

I’d posit there are some very simple litmus tests any government agency can put against their projects to determine if they are truly opening data:

1) Is there a download button?

In order for any data to be open you need to be able to download the data so that you can remix, reuse and share the data. Data and the government agency that supplies it are not transparent if you can’t download the raw data. PDF’s and web services don’t count. They can be useful additions to the raw data, but they are not a replacements.

2) Data should always be linked to the derivative works created with it.

Take Geoplatform.gov – a nice mash up of data from Data.gov but I have no way to get back to the data to re-purpose it in a useful way other than a few links to PDF’s – the scourge of open data.

You’ve highlighted great open data, now let the community get to it easily. It is like running a great advertisement and then giving your consumer no easy way to buy the product. All government mashups – map and otherwise should have mandatory link back to the raw data for download.

3) Downloading should never be more than two clicks away (ideally one).

Whether it is Geodata.gov or any other government site you should always be able to quickly and clearly get to raw data download easily. It does not matter whether you are a GIS user or Joe Six-pack finding the content you want should be a simple process. In the recent open data discussions folks have been holding up the features of Geodata.gov as the standard, but I think we can do better:

Five clicks (after you find the advanced search) and navigating a lot of text is a tough task even for a GIS professional.

4) When you get to the data it should be available in multiple formats and be easy to use (i.e. data dictionaries).

Users come in a lot flavors whether they are geo or non-geo. They should have a selection of data formats to work with. There are lots of great file format transformation tools out there both open source and licensed. While easy to use is subjective. Let’s start with demonstrating how our current options are not easy:

Following the last example I did a search for demographics on Geodata.gov and got a result for “200 county demographics”. I received an option to download a PDF or a zip file. PDF is where data goes to die, so obviously I went for the zip file. I unzipped the file and found a .csv. Cool but not really geodata. All the same I opened it up to find lots of data but no attributes for the columns. Annoying, but no big deal I’ll shoot back to the metadata – the essence of Geodata.gov. There is a list of attributes and definitions but no way at all to map those back to the spreadsheet. Rendering both the spreadsheet and the metadata useless.

5) The data should be searchable and portable.

It was great to get the clarification that ESRI will be implementing the OGC CSW standard. A standard is always better than a proprietary format or approach. Folks may disagree but I’ll go a step further and say a Web aligned standard is most often better than a GIS standard. The CSW spec took us 14 days to implement and Opensearch took a little over a day. The CSW spec is over 200 pages and the OpenSearch spec is a couple of pages. Good news being the OGC is considering adopting OpenSearch (only 24 pages . In our experience Opensearch has allowed the community to create a large number of simple interfaces into Geocommons – including ArcMap, Google Earth, Jive, Windows 7, Chrome, IE8 and Firefox.

There are a lot of other great things that can be added on top of these open data basics, but they are not substitutes. Web services and API’s are all great things, but in my mind they come after the basic data access. Creating sophisticated mash-ups and dashboards with data are also very useful, but if the data is not also directly available (a download button) fail the litmus test.

There is a ton of potential in opening up government data for better transparency, greater innovation and sustained economic growth. In order to capture that potential though we must truly open data, and not let the idea be subverted or hijacked by special interests.

Lastly, after revisiting Geodata.gov today and trying to actually work with it I can only come to one conclusion. If we truly care about opening up government geo-data – rebuild the whole thing from the ground up. I’m sure it was great in 2004, but the Web and GIS has moved on.

Popularity: 13% [?]

At FortiusOne we have a history of mapping elections. Previously in this blog Sean Gorman discussed “Leveraging Open Data for Transparency” during the 2009 Afghanistan elections. This blog also looked at rates of violence and its potential impact on voter turnout in Afghanistan as well. We also provided a dashboard for the Virginia election as well, looking at voter turnout and the Republican victories.

When VoteGeorgia.ge decided to use GeoCommons’ maps to display election monitoring information for the Georgian elections we were excited to be able to help. The idea was to provide all the election data all in one place so it could be accessible to all. Using the sharing features built into GeoCommons maps were provided for the site.

One thing that was going to be important during the election period was frequent updating of the map.  This is why feeds were used to keep the various data overlays up to date.  Using the “Add a URL Link” feature in Finder! the GeoRSS feeds were added.  URL linking works for KML and CSV files as well.  How often a feed is updated is dependent on the update rate of your individual dataset.

With the datasets being updated then the next step was to get maps in the VoteGeorgia website.  After a map was made in Maker! the embed code for that map can be embedded in another website.  To embed maps in your own website click the “Details” button on a map and then click “Embed this map in your website.”  A textbox with an embed code will appear, just copy and paste this into the HTML of the site where you want to share the map.

To see the results of this simple integration with GeoCommons go to VoteGeorgia.ge.

Popularity: 10% [?]

Today I was on a panel at the InterAction Forum, which is a conference for member organizations within InterAction.  InterAction is an coalition of NGOs with over 180 member organizations.   In the past few months at FortiusOne we’ve been assisting them in tracking their projects and sharing their geographic data. The panel I participated in was titled “Mapping the Haiti Earthquake: Open Geospatial Data and Transnational Social Collaboration.”  It was moderated by Andrew Schroeder of Direct Relief International with the other participants being Christiann Adams from Google, Kimberly Konkel from the Department of Health and Human Services, Nathan Heard of the State Department and Salim Sawaya from ESRI.

One of the points the panel wanted to highlight was the collaboration after the Haiti earthquake between corporations, government institutions and newer participants such as Ushahidi and OpenStreetMap.  As someone who participated early on in the collaboration it was interesting to finally meet people who previous I only knew through phone calls and emails.  One of the questions coming out was is Haiti special or will this happen again?  Was there something about the particular incident or has something fundamentally changed?

I think there are multiple sides to this question.  There is the technology, the people side, the geographic side and the sheer magnitude of the disaster.  I think as far as technology goes things have changed, people can find collaborators in a way that wasn’t possible before.  By finding collaborators individuals were able to make a difference far away from the scene of the actual crisis.  I think the people side is more complex, there was something compelling in the United States with Haiti being so close.  Though now people have formed powerful networks and know that when others need help there is often something they can do.  As far a geography the proximity of Haiti to the United States I believe made the government and its citizens respond in a way they haven’t before.  Then there was the size of the disaster which compelled many people to help.

I think all these factors have changed how people will respond in to a crisis in the future.  Now when an earthquake or flood hits I immediately see the various networks activate, seeing if they can do anything to help.  In some cases such as with the oil spill there is something the technology community can do.  In others there are other factors that either prevent people from helping or the help is unneeded.

Response also will change as groups continue to plan better for the future.  One example of this is the CrisisCamp and the Random Hack of Kindness this weekend.   Another was the Understanding Risk Conference at the World Bank earlier in the week.  I’m still participating in capacity building for OpenStreetMap in Haiti with the Humanitarian OpenStreetMap Team.  Data will continue to be available on GeoCommons when other events occur such as the oil spill in the Gulf of Mexico.  And technology will continue to improve making collaboration easier.

Popularity: 19% [?]