azavea / acs-alchemist Goto Github PK

Many of the license files included with the install appear to have Unix newlines (e.g. everything in log4net directory). This isn't a problem per-se, but most people will be reading them w/ Notepad, which doesn't grok Unix, so they look kinda strange.

to force the user to make a conscious choice

revise the config generator to support 2011 data
add a config file to what is shipped
test 2011 data release to make sure Census Bureau file structure has remained stable.

Readme that pops up after install has no references to GUI version

I suspect this might be an issue regarding our SRID file using "US Survey Foot" instead of "Foot_US" -- fix in place, waiting to see if it is resolved.

In the config file (for 2010), the counties URL and file name lines should look like this:

"ShapeFileCountiesURL": "http://www2.census.gov/geo/tiger/TIGER2010/COUNTY/2010/",
"ShapeFileCountiesFilename": "tl_2010_{FIPS-code}_county10.zip",

Additional testing will likely be necessary.

this was intentionally made gray so it wouldn't be distracting, but we can change it.

need more details / description for this feature

Alchemist has recently encountered issues, and will no longer operate. For example, when running a query for total population at the Pennsylvania census block group level for 2014, the following error messages appear:
When downloading the column mappings files -
Could not download file at http://www2.census.gov/programs-surveys/acs/summary_file/2014/data/2014_5yr_Summary_FileTemplates.zip

When downloading the census data -
Could not download file at http://www2.census.gov/programs-surveys/acs/summary_file/2014/data/5_year_by_state/Pennsylvania_Tracts_Block_Groups_Only.zip

Perhaps this may be due to census data sites changing? These same queries have worked on 3/24/17 with no issues. Any help would be greatly appreciated!

either prevent multiple instances from running, or prevent them from interfering with eachother

File -> New Job doesn't ask if you want to save your current job first
Exiting the app doesn't ask if you want to save your current job first

Potentially:

• File/Personal GeoDatabase
• .xls / .csv file
• etc

some error about being unable to find a particular tract / block group.. Something about the duplicate logrecno fix, or polygon merge fix? Or possibly a result of mis-fixed-formatting of tract numbers?

It turns out some of the CENSUS_TABLE_IDs (e.g. B05002013) have multiple sequence numbers associated with them. When we do our SQL query against the table, it returns all the rows with those IDs, resulting in duplicate columns in the logic.

Because the columns have the same name, only one column is ever created so this isn't a huge problem, but we should probably fix it. We can either modify our query statement, or add a unique constraint to the columnMappings table and skip entries that would violate it during creation.

--Is this still an issue? Can we get steps to reproduce it?

File -> New Job doesn't clear progress bar at bottom, or Current Job File field at top

Possible solutions:
a) write our own implementation of AttributesTable (it has an interface wrapper) that uses less memory
b) dump everything into a table and use spatialite's .dumpshp command

"It would be nice to have tooltips for the labels, e.g. hovering over "4. Variable File" would briefly describe what a variable file is"