Data.gov: Evolving Data.gov With You

• Data.GOV
• Chief Information Officers Council

michael.daconta

Member since 3 years ago

Badges

• Newbie

0 ideas posted

0 comments 0 votes

3 points View the Leaderboard

First name:

Michael

Last name:

Daconta

Username:

mdaconta

michael.daconta is a member of:

• Evolving Data.gov with You

Activity Stream

Ideas contributed by michael.daconta

• 1 vote

  This idea is closed.
  Solution Architecture »

  Publish Standards on data.gov for Supported Data Formats

  In examining an interesting data set that is supplied as a Comma-Separated Values (CSV) file, I noticed that the CSV file does not have the header line. It is customary for a CSV file to have the first line be the column names separated by commas. This dataset does not have the header line: http://www.data.gov/details/1617 So, what is the solution here? I recommend that for each format supported on data.gov there exists ...more »

  In examining an interesting data set that is supplied as a Comma-Separated Values (CSV) file, I noticed that the CSV file does not have the header line. It is customary for a CSV file to have the first line be the column names separated by commas.

  This dataset does not have the header line:

  http://www.data.gov/details/1617

  So, what is the solution here? I recommend that for each format supported on data.gov there exists an endorsed standard (existing or newly created) for how that should be formatted by agencies.

  So, in this case, data.gov would post on their site a standard for how agencies should produce CSV files and hopefully that would include a recommendation that each CSV file produced begins with a header line that has the column names separated by columns for the data.

  (of course, each column name should have a definition in the data dictionary as proposed in a different idea).

  « less full details »

  1 comment for Idea: Publish Standards on data.gov for Supported Data Formats

  Submitted by michael.daconta 3 years ago

• I agree

  7 votes

  I disagree
  This idea is active.
  Solution Architecture »

  Remove and Guard Against (Validate) "Junk" Records

  I tried to download a dataset that interested me called the "Occupational Outlook Handbook". The data.gov "record" says that it is a CSV dataset. However, when you click the csv link you go to a website that does not allow you to download the dataset! The data.gov link is: http://www.data.gov/details/336 This is a failure of the simplest validation possible - a link that is supposed to be to a dataset must be tested ...more »

  I tried to download a dataset that interested me called the "Occupational Outlook Handbook". The data.gov "record" says that it is a CSV dataset. However, when you click the csv link you go to a website that does not allow you to download the dataset!

  The data.gov link is: http://www.data.gov/details/336

  This is a failure of the simplest validation possible - a link that is supposed to be to a dataset must be tested that it actually points to the type of dataset advertised! Otherwise, this is a violation of trust and a large waste of time. How many people will be interested in this dataset, spend 10-15 minutes of their day to access it, only to walk away disappointed. Hmmmm... 10-15 minutes wasted X millions of people = fail!

  These types of "junk" or misleading records should be removed from data.gov. Additionally, these types of junk records must be prevented from being added to the data.gov catalog.

  « less full details »

  1 comment for Idea: Remove and Guard Against (Validate) "Junk" Records

  Submitted by michael.daconta 3 years ago

• I agree

  10 votes

  I disagree
  This idea is active.
  Solution Architecture »

  Don't Inflate Counts! (aka let a Dataset contain multiple files)

  When searching for interesting datasets it can be very frustrating to see identical datasets (except by year) cluttering the search results. To me, this seems like just an attempt to inflate the counts by treating separate years or geography as separate datasets. Here is my proposed simple rule: If the dataset metadata fields (i.e. like "Coverage Date") are the same, it is the SAME dataset. Just because the instance ...more »

  When searching for interesting datasets it can be very frustrating to see identical datasets (except by year) cluttering the search results. To me, this seems like just an attempt to inflate the counts by treating separate years or geography as separate datasets.

  Here is my proposed simple rule:

  If the dataset metadata fields (i.e. like "Coverage Date") are the same, it is the SAME dataset. Just because the instance data varies by year or geography, that should be accounted for by allowing a dataset to contain multiple files.

  « less full details »

  3 comments for Idea: Don't Inflate Counts! (aka let a Dataset contain multiple files)

  Submitted by michael.daconta 3 years ago

• I agree

  22 votes

  I disagree
  This idea is active.
  General »

  Transparency of the Process behind the Data

  A citizen should be able to request a description of how a data element is derived. For example, the very controversial - "saved or created jobs" data element should have an explanation of the formula or the derived fields used to calculate the number. Another example, would be the all important unemployment data - how is that percentage calculated? This would give the public more confidence in using the data. In ...more »

  A citizen should be able to request a description of how a data element is derived. For example, the very controversial - "saved or created jobs" data element should have an explanation of the formula or the derived fields used to calculate the number. Another example, would be the all important unemployment data - how is that percentage calculated? This would give the public more confidence in using the data.

  In summary, the agencies need to be able to describe, in detail, how they collect or create the data they are providing to the public. Additionally, to what professional standards they are adhering to assure the quality and trustworthiness of that data.

  « less full details »

  Add your comment for Idea: Transparency of the Process behind the Data

  Submitted by michael.daconta 3 years ago

• I agree

  17 votes

  I disagree
  This idea is active.
  Strategic Intent »

  Create an International data.gov working group

  In light of the excellent work other governments are doing in this area, the US should spearhead an international group (maybe in conjunction with the UN) to share best practices, cross-fertilize experience and maybe even share common code or data formats.

  Add your comment for Idea: Create an International data.gov working group

  Submitted by michael.daconta 3 years ago

• I agree

  24 votes

  I disagree
  This idea is active.
  Solution Architecture »

  Improve Browsing of Data.gov

  While there are quite a few discussions on this site about search; I would like to see improved browsing via a robust taxonomy/folksonomy of topic areas. In fact, I would recommend both a top-down/bottom-up approach where you begin with a top-down taxonomy but allow it to be extended via topic area suggestions and popular keywords. As a citizen, I don't always know what I want but want to browse and see what is available. ...more »

  While there are quite a few discussions on this site about search; I would like to see improved browsing via a robust taxonomy/folksonomy of topic areas. In fact, I would recommend both a top-down/bottom-up approach where you begin with a top-down taxonomy but allow it to be extended via topic area suggestions and popular keywords.

  As a citizen, I don't always know what I want but want to browse and see what is available.

  « less full details »

  5 comments for Idea: Improve Browsing of Data.gov

  Submitted by michael.daconta 3 years ago

• I agree

  29 votes

  I disagree
  This idea is active.
  Key Partnerships »

  Public Data.gov Roadmap

  Many high-profile projects that rely on a community of supporters, like mozilla.org and open office, post a public roadmap that details major forthcoming plans and milestones. OMB/GSA should do the same for data.gov so that the public can see when planned innovations will occur and that they are actually influencing the evolution.

  Add your comment for Idea: Public Data.gov Roadmap

  Submitted by michael.daconta 3 years ago

• I agree

  26 votes

  I disagree
  This idea is active.
  Developer's Corner »

  Developer Wiki

  Though there is another idea regarding a wikipedia entry... I think it is important to have a separate wiki set up for developers as there are many developer topics where collaboration is key. Of course, there are tons of precedent for this whereby many major development efforts/sites have developer wikis.

  7 comments for Idea: Developer Wiki

  Submitted by michael.daconta 3 years ago

• I agree

  13 votes

  I disagree
  This idea is active.
  General »

  Top 5 metadata elements?

  What would the community feel is the top 5 metadata elements that should be mandatory for a dataset catalog record AND of those 5 which do you think is most important to developers and citizens wanting to discover and use datasets that are relevant to them? Here is a recommendation: 1. Subject Coverage (aka topic or "WHAT", for a hierarchical tree display/browsing) 2. Schema Location (for structured datasets) 3. Geospatial ...more »

  What would the community feel is the top 5 metadata elements that should be mandatory for a dataset catalog record AND of those 5 which do you think is most important to developers and citizens wanting to discover and use datasets that are relevant to them?

  Here is a recommendation:

  1. Subject Coverage (aka topic or "WHAT", for a hierarchical tree display/browsing)

  2. Schema Location (for structured datasets)

  3. Geospatial Coverage (aka "WHERE", for a map display)

  4. Temporal Coverage (aka "WHEN", for a timeline display)

  5. Biographical Coverage (aka "WHO", for a link analysis display)

  Note: by "Coverage" I mean (like DDMS) the contents of the dataset and not metadata about the publisher (though that is important too, just not in regarding to relevance).

  What are your Top 5?

  « less full details »

  9 comments for Idea: Top 5 metadata elements?

  Submitted by michael.daconta 3 years ago

• I agree

  14 votes

  I disagree
  This idea is active.
  Agency Next Steps »

  Quality controls/checklist

  The agency posting the data should assert that it has performed certain quality gates/checks on the datasets it is publishing. I just downloaded one dataset provided as a spreadsheet and one of the columns is: "Contructor" Either that is a weird term or it is an error and meant to say "Contractor". Looking at the data (which is a bunch of acronyms) it is probably the latter. Did whoever published the dataset check ...more »

  The agency posting the data should assert that it has performed certain quality gates/checks on the datasets it is publishing.

  I just downloaded one dataset provided as a spreadsheet and one of the columns is:

  "Contructor"

  Either that is a weird term or it is an error and meant to say "Contractor". Looking at the data (which is a bunch of acronyms) it is probably the latter.

  Did whoever published the dataset check the metadata?

  Did they insure all definitions were published?

  Have the data values been profiled?

  If any columns are codes, have those been published?

  Have constraints on values been published?

  Are business rules on the values published?

  On the catalog entry their should be a data quality controls checklist that is completed by the submitter before publishing the data.

  That checklist should correspond to the Agency internal governance procedures.

  « less full details »

  5 comments for Idea: Quality controls/checklist

  Submitted by michael.daconta 3 years ago

• I agree

  73 votes

  I disagree
  This idea is active.
  Solution Architecture »

  Definitions for every column

  I just downloaded some energy data from data.gov on nuclear reactors and one of the columns is:

  "NRC Unit"

  I have no idea what that means? Every column or field of data should have a definition and that should be available on data.gov or in a standard format with the dataset. In this case, the data dictionary field that the catalog record links to does not have the definition of this field.

  19 comments for Idea: Definitions for every column

  Submitted by michael.daconta 3 years ago

• I agree

  12 votes

  I disagree
  This idea is active.
  General »

  Data.gov Accountability Dashboard on Front Page

  The open government directive has specific requirements like the following: Within 45 days, each agency shall identify and publish online in an open format at least three high-value data sets ... and register those data sets via data.gov... These must be data sets not previously available online or in a downloadable format. Tracking those specific requirements on data.gov (separate from the open government dashboard ...more »

  The open government directive has specific requirements like the following:

  Within 45 days, each agency shall identify and publish online in an open format at least three high-value data sets ... and register those data sets via data.gov... These must be data sets not previously available online or in a downloadable format.

  Tracking those specific requirements on data.gov (separate from the open government dashboard which is broader) will help data.gov adoption by improving the quality and relevance of data available to developers and the public.

  Besides agency compliance, it would be good to post a ranking of the most requested data on the dashboard.

  « less full details »

  Add your comment for Idea: Data.gov Accountability Dashboard on Front Page

  Submitted by michael.daconta 3 years ago

• I agree

  6 votes

  I disagree
  This idea is active.
  Developer's Corner »

  Existing Catalog APIs - Developer Favorites or Standard API?

  Like many commercial product catalogs (i.e. Amazon.com) there is a web services api to search/access the catalog. Of course there are also REST APIs to do this (don't want a REST versus web services flame war here).

  What catalog API would developers suggest for data.gov?

  Is there a standard catalog API suitable for data.gov?

  Discuss via your comments...

  4 comments for Idea: Existing Catalog APIs - Developer Favorites or Standard API?

  Submitted by michael.daconta 3 years ago

• I agree

  19 votes

  I disagree
  This idea is active.
  Developer's Corner »

  Semantic.data.gov use cases needed.

  As discussed in the CONOPS, semantic.data.gov will be an adjunct, experimental site to assist in the evolution of data.gov towards greater semantics. This site will also learn from lessons learned via data.gov.uk which is also exploiting semantic technologies. What would be the most important initial use cases the semantic web community would like to see semantic.data.gov tackle? Ontology development? Rule based alerts? ...more »

  As discussed in the CONOPS, semantic.data.gov will be an adjunct, experimental site to assist in the evolution of data.gov towards greater semantics. This site will also learn from lessons learned via data.gov.uk which is also exploiting semantic technologies.

  What would be the most important initial use cases the semantic web community would like to see semantic.data.gov tackle?

  Ontology development?

  Rule based alerts?

  Reasoning use cases?

  It is important to select the top use cases that demonstrate the power of semantic web techniques and that can be practically implemented.

  « less full details »

  6 comments for Idea: Semantic.data.gov use cases needed.

  Submitted by michael.daconta 3 years ago

Displaying all 14 Ideas