How do I find a place on the map?

There are several ways to navigate around the map. You can use the navigation bar like you see to the left of the page to zoom in and out on the map. The round circle allows you to pan right, left, up and down around the map. The other way to navigate the map is to use the ‘Find Address’ box on the top left hand side to enter your address or a latitude and longitude and Bing maps will geocode this address and take you as close as possible to the point on the map and place a push pin on the spot where Bing Maps thinks the address is.

Turning On Layers

On the left hand side is the layer control. With this tool you can check and uncheck the different layers that will turn the layers on and off. On the example to the left I have the ‘Parcels’ the ‘Census Roads’ and the ‘Watch List’ checked or turned on and ‘Census Blocks’ unchecked or turned off. The little question marks ‘?’ can be clicked to give help about the area near the question mark.

You can see that the borders around the numbers and check box are in a deep green color. This is to show that at your current zoom level these layers would show if turned on. The two numbers (zoom range) that you see will tell you between what zoom levels the layer will appear on the map. If you want to see a layer closer in or further out you can change these numbers. If you show detailed information for too big of an area then the map will take longer to load.

You can see what zoom level you are on by looking at the top right hand corner of the screen. The zoom level of the graphic shown here is at ‘18’. This would show all the layers checked above. You will also notice the words ‘Mouse Lat/Lon: 30.4389/-84.2809’. If you click your mouse anywhere on the map the program will display the latitude and longitude of the point you click on the map. The color surrounding the layer name is the same color you will see on the map surrounding the different parts of geography.

Displaying Data

As you move your mouse over different pieces of geography you will see the block will be highlighted. At the same time on the left hand side of the screen you will see different information displayed for each block. To the left you can see part of the information for one parcel from the 2009 parcel data.

Road Data -You can see on the right an example of road data. The section of road that is in focus is highlighted when you move the mouse over it.  The data fields are as follows:

countyfp	Index of the data.  073 is a census number assigned to Leon County.
tlid	Census tiger line id number that identifies the road segment.
mtfcc	A census line segment type number.
fullname	What the name is in the census edges file
Streets	The different names the census has for the road segment.
Numbers	Gives all the different address ranges the census has for the road segment.
Zips	The zip codes for the road segment. Sometimes one side of the road will have a different zip then the other.

Hiding Windows - Notice the little down arrow next to the word ‘Data’ in the upper left hand corner of the graphic above. If you click on this it will hide the data layer. You can also hide the layers window or if you click on the ‘Hide Menu’ on the upper left hand corner of the map screen you can hide all the menus and see the map with a full screen.

Finding Where People May Be Missed – By comparing the data found in the census road file to the address in the parcel file you can look for missing or miss aligned streets, streets with different names than the census, or maybe a street has no name at all in the census data. You can look for parcels that have an address that does not fall in the range of addresses from the census data. Some places you will find that the census will have one side of the street but not the new houses on the other.

Watch List

In October we ran a program over the state looking for places where the census geography did not seem to match with the parcel geography. The red blotches are example areas in Central Florida that we have identified as places the census may miss sending census forms to. Since we ran these programs the census has gone through the LUCA challenge project, the new construction project and has worked with the State of Florida to try to take care of some of these problems. We suspect that there may still be problems with new construction where the aerials do not show houses.

The above watch list display is using the satellite view. You may find it easier to see using the road view. To switch to road view, find the navigation bar at the top left of the map and click "Road". The map view will switch to the road mode.

This view may make it a little easier to see the watch list items. Note: Viewing at broader zoom levels with the watch list turned on may cause large time delays as the map is updated.

On the graphic above you can see some examples of what to look for to find people that the census may miss. The blue lines are the census roads. You can see some match well and some do not match at all. The green lines are the parcels. The small red boxes are the areas where our program discovered possible problems. Then you can see areas where Bing shows there is a road but there is no blue census road and no red blocks. These are areas to see if the people received their census forms. If you live in one of these areas please email and tell us you were not counted at feedback@myfloridacensus.gov and after April 12th come back to the map and enter your address to flag this area for the census.

Using the Census Geography

With the "Block" layer selected, you can mouse over the census block geography to highlight any census block and display its data. When you do this you will see the census block number, the census housing units and the census living quarters numbers appear on the data window. You can examine the parcel data for each parcel and count how many housing units the parcel data tells you there are. On the census block above there were 26 housing units and 6 group quarter units. By looking at the parcel data we found there was a private school. If I had local knowledge I may know if there was provision for students to live on the property. By counting the properties and looking at the property types I can check the October 2009 census address counts.

How the data was obtained.

The census data found on the website came from the census submission to the Florida Legislature for the purpose of verifying the VTD's for Florida. VTD stands for Voter Tabulation Districts. VTDs are the building blocks that future voting precincts can be built on. This data is the last set of data we expect to be released before the final 2010 tiger census data.

Census data can be divided into several types, including county, tract, block group and block data. It is by these geographic areas that the results of the census information will be released to the public.  These types are displayed as separate layers on the map.

As you cursor over the data in these layers you will see several numbers displayed on the left side in the data window.   They include:

 

Census Housing Units	This number comes from the October LUCA (Local Update of Census Addresses) verification data. This is the number of addresses that the census has for each area. This number will be from the New Construction Program and from addresses that were originally rejected in the LUCA process but were later challenged by the local governments and approved by the Census upon further review.
Census Living Quarters	The census group quarters number. Group quarters are places like school dorms, group homes, nursing homes etc. that have a number of unrelated people living at the same address

 

The next geography type that we are displaying is the place or city data and the unincorporated area data. With this data you can see where the borders of cities and other places are.

Another type of census geography is the census road file.  This is a subset of the census 'edges' file with all but the road segments deleted.  The road layer has the following data items:

Streets	This has all the different names that the census has for each road segment from the ALLNAMES file
address ranges	This is found in the census ADDR file for each road segment with a ‘R’ or ‘L’ after it to say whether the numbers are on the right or left side of the road.

 

Parcels – The other major data set is the statewide parcel file. This started with the statewide 2009 parcel shapefiles found on the Florida Department of Revenue (DOR) FTP site. This data was run through conversion programs to convert all this shape data to Well Known Text (WKT) format. Any shape that we found corrupted or that did not have a ‘MPID’ DOR id number was not included in the final set.

The information found with the parcel is from 2009. Up-to-date information can be obtained at any county property appraiser website. The purpose of the parcel data on our website is to let anyone compare the census data to the property data to the Bing Maps’ data and be able to easily determine where the census has missed roads, miss aligned roads, has the wrong name for a road or does not have the correct address ranges for a road.

Some data fields for parcels are:

Address	This is the address of the property after being standardized by the postal CASS software.
Property Use	A short description of how the property is used.
Units	The number of residential units the property appraiser said was on the parcel.
Buildings	This is the number of buildings on the parcel.
Year Built	The year the property was improved.
Square Foot	The square footage of the building on the property.
Total Value	The value of the property as assessed by the county appraiser.
Owner	The owner of record in 2009.
Land Value	The value of the land.
Land Sq ft	The square footage of the parcel.
Legal	A short legal description.
Tax Authority	The tax authority code. You may be able to compare this to the census city data to see if they have the city lines correct.
Homestead	The amount of homestead or other exemption the property receives.
Parcel ID	The state parcel id number.

 

How we did it

Census Geography – We took census geography and used a program tool in PostGIS plugins to import the census shape files into postgresql. We found this tool to be the fastest and the best tool of several that we used for importing the spatial data. We then used PostGIS to translate the geographic data into Well Known Text (WKT). This data we then imported in to SQL Server 2008. We used Spatial commands MakeValid() and STIsValid() in this tool to check the data and to fix data that was found to be incorrect. We then made a simplified version of all but the block and the road data to be used in the program with the .Reduce(.001)command in SQL Server. Some of the WKT definitions were bigger than 1 MB.

For the road data we wrote a program to combine all the names and address ranges for each road segment into one field that we could easily display.

The Parcel Data – This data was the biggest challenge. This data came from 67 different property appraisers and tended to be of very different qualities. Some of the files we had trouble converting to WKT even though we used four different tools. For some of the counties we could not convert directly to WKT but found a tool that converted it to KML format and then wrote a tool to convert it from there to WKT.

The final data set was just under 9 million records. Another problem with this data was that we had thought that the MPID field was a unique identifier for polygon to tax record. If a tax record had several segments it was a multi polygon. We found that there were about 400 thousand records that had multiple polygon records with the same MPID. These polygons were not used in the final dataset because of time and should be combined at a later date.

At this point we know that there are 4 counties, Pinellas, Jackson, Charlotte and Clay were the projection is slightly off. Many counties have some missing parcel polygons and some polygons are missing data. As time goes on these problems will be corrected.

 

The Application

The application is a heavy mixture of Microsoft technology. 

 

It is hosted in Microsoft's Windows Azure cloud and runs in your web browser using Microsoft Silverlight for cross browser compatibility. Silverlight is based on the .NET application framework, an our application is written in C#.  Within Silverlight we use the Microsoft Bing Maps control to provide the map interface.  We access our Microsoft SQL Azure database through the Microsoft Windows Communciation Foundation communications protocol.  Within Windows Azure we are using stored procedures writtin in Structured Query Language (SQL).  Geocoding (Taking an address and translating it to Latitude/Longitude coordinates) is done using Microsoft Geocoding web service.  The Bing Map control accesses the Bing Maps service to provide the map tiles.

Hosting in the Cloud – We picked this way of hosting the application because we knew that the use of the application would be high for a few critical weeks and much lower the rest of the time. The Windows Azure hosting gives you the ability to expand to more servers and bandwidth in a matter of minutes.  We will monitor the application to see what we need to do to handle the load as time goes on.

 SQL Azure Database – For storage we used two 10 GB Azure SQL databases.  As we build our redistricting application we plan to use Microsoft Azure table storage which we think will give us the storage needed for the redistricting application we will be developing. 

Spatial data – The big drawback to Azure storage for us was that it does not handle the SQL Server spatial extensions found in SQL Server 2008. The other hosting services we investigated could provide SQL Server but would only provide 1 to 2 GB of database storage. Not near enough for our project.

To overcome this drawback we created our own spatial indexing system. This system divided the state into a set of different size grids depending on the density of the data and then with a math formula determined which grid each polygon was part of based on the maximum and minimum latitude and longitude found in each polygon. We then created an index to the grid boxes based on this information.

When we want to display a screen of polygons we determine with a math formula which grid boxes are needed to display the screen and return a distinct set of polygons.

We hope this description of the MyFloridaCensus.gov will help you in your use of this program. If you have any suggestions or comments please feel free to send them to feedback@myfloridaCensus.gov